We are excited to bring Transform 2022 back in-person July 19 and virtually July 20 - 28. Join AI and data leaders for insightful talks and exciting networking opportunities. Register today!

The words artificial intelligence (AI) have been used to describe the workings of computers for decades, but the precise meaning shifted with time. Today, AI describes efforts to teach computers to imitate a humans ability to solve problems and make connections based on insight, understanding and intuition.

Artificial intelligence usually encompasses the growing body of work in technology on the cutting edge that aims to train the technology to accurately imitate or in some cases exceed the capabilities of humans.

Older algorithms, when they grow commonplace, tend to be pushed out of the tent. For instance, transcribing human voices into words was once an active area of research for scientists exploring artificial intelligence. Now it is a common feature embedded in phones, cars and appliances and it isnt described with the term as often.

Today, AI is often applied to several areas of research:

There is a wide range of practical applicability to artificial intelligence work. Some chores are well-understood and the algorithms for solving them are already well-developed and rendered in software. They may be far from perfect, but the application is well-defined. Finding the best route for a trip, for instance, is now widely available via navigation applications in cars and on smartphones.

Other areas are more philosophical. Science fiction authors have been writing about computers developing human-like attitudes and emotions for decades, and some AI researchers have been exploring this possibility. While machines are increasingly able to work autonomously, general questions of sentience, awareness or self-awareness remain open and without a definite answer.

[Related: Sentient artificial intelligence: Have we reached peak AI hype?]

AI researchers often speak of a hierarchy of capability and awareness. The directed tasks at the bottom are often called narrow AI or reactive AI. These algorithms can solve well-defined problems, sometimes without much direction from humans. Many of the applied AI packages fall into this category.

The notion of general AI or self-directed AI applies to software that could think like a human and initiate plans outside of a well-defined framework. There are no good examples of this level of AI at this time, although some developers sometimes like to suggest that their tools are beginning to exhibit some of this independence.

Beyond this is the idea of super AI, a package that can outperform humans in reasoning and initiative. These are largely discussed hypothetically by advanced researchers and science fiction authors.

In the last decade, many ideas from the AI laboratory have found homes in commercial products. As the AI industry has emerged, many of the leading technology companies have assembled AI products through a mixture of acquisitions and internal development. These products offer a wide range of solutions, and many businesses are experimenting with using them to solve problems for themselves and their customers.

Leading companies have invested heavily in AI and developed a wide range of products aimed at both developers and end users. Their product lines are increasingly diverse as the companies experiment with different tiers of solutions to a wide range of applied problems. Some are more polished and aimed at the casual computer user. Others are aimed at other programmers who will integrate the AI into their own software to enhance it. The largest companies all offer dozens of products now and its hard to summarize their increasingly varied options.

IBM has long been one of the leaders in AI research. Its AI-based competitor in the TV game Jeopardy, Watson, helped ignite the recent interest in AI when it beat humans in 2011 demonstrating how adept the software could be at handling more general questions posed in human language.

Since then, IBM has built a broad collection of applied AI algorithms under the Watson brand name that can automate decisions in a wide range of business applications like risk management, compliance, business workflow and devops. These solutions rely upon a mixture of natural language processing and machine learning to create models that can either make production decisions or watch for anomalies. In one case study of its applications, for instance, the IBM Safer Payments product prevented $115 million worth of credit card fraud.

Another example, Microsofts AI platform offers a wide range of algorithms, both as products and services available through Azure. The company also targets machine learning and computer vision applications and like to highlight how their tools search for secrets inside extremely large data sets. Its Megatron-Turing Natural Language Generation model (MT-NLG), for instance, has 530 billion parameters to model the nuances of human communication. Microsoft is also working on helping businesses processes shift from being automated to becoming autonomous by adding more intelligence to handle decision-making. Its autonomous packages are, for instance, being applied to both the narrow problems of keeping assembly lines running smoothly and the wider challenges of navigating drones.

Google developed a strong collection of machine learning and computer vision algorithms that it uses for both internal projects indexing the web while also reselling the services through their cloud platform. It has pioneered some of the most popular open-source machine learning platforms like TensorFlow and also built custom hardware for speeding up training models on large data sets. Googles Vertex AI product, for instance, automates much of the work of turning a data set into a working model that can then be deployed. The company also offers a number of pretrained models for common tasks like optical character recognition or conversational AI that might be used for an automated customer service agent.

In addition, Amazon also uses a collection of AI routines internally in its retail website, while marketing the same backend tools to AWS users. Products like Personalize are optimized for offering customers personalized recommendations on products. Rekognitition offers predeveloped machine vision algorithms for content moderation, facial recognition and text detection and conversion. These algorithms also have a prebuilt collection of models of well-known celebrities, a useful tool for media companies. Developers who want to create and train their own models can also turn to products like SageMaker which automates much of the workload for business analysts and data scientists.

Facebook also uses artificial intelligence to help manage the endless stream of images and text posts. Algorithms for computer vision classify uploaded images, and text algorithms analyze the words in status updates. While the company maintains a strong research team, the company does not actively offer standalone products for others to use. It does share a number of open-source projects like NeuralProphet, a framework for decision-making.

Additionally, Oracle is integrating some of the most popular open-source tools like Pytorch and Tensorflow into their data storage hierarchy to make it easier and faster to turn information stored in Oracle databases into working models. They also offer a collection of prebuilt AI tools with models for tackling common challenges like anomaly detection or natural language processing.

New AI companies tend to be focused on one particular task, where applied algorithms and a determined focus will produce something transformative. For instance, a wide-reaching current challenge is producing self-driving cars. Startups like Waymo, Pony AI, Cruise Automation and Argo are four major startups with significant funding who are building the software and sensor systems that will allow cars to navigate themselves through the streets. The algorithms involve a mixture of machine learning, computer vision, and planning.

Many startups are applying similar algorithms to more limited or predictable domains like warehouse or industrial plants. Companies like Nuro, Bright Machines and Fetch are just some of the many that want to automate warehouses and industrial spaces. Fetch also wants to apply machine vision and planning algorithms to take on repetitive tasks.

A substantial number of startups are also targeting jobs that are either dangerous to humans or impossible for them to do. Against this backdrop, Hydromea is building autonomous underwater drones that can track submerged assets like oil rigs or mining tools. Another company, Solinus, makes robots for inspecting narrow pipes.

Many startups are also working in digital domains, in part because the area is a natural habitat for algorithms, since the data is already in digital form. There are dozens of companies, for instance, working to simplify and automate routine tasks that are part of the digital workflow for companies. This area, sometimes called robotic process automation (RPA), rarely involves physical robots because it works with digital paperwork or chit. However, it is a popular way for companies to integrate basic AI routines into their software stack. Good RPA platforms, for example, often use optical character recognition and natural language processing to make sense of uploaded forms in order to simplify the office workload.

Many companies also depend upon open-source software projects with broad participation. Projects like Tensorflow or PyTorch are used throughout research and development organizations in universities and industrial laboratories. Some projects like DeepDetect, a tool for deep learning and decision-making, are also spawning companies that offer mixtures of support and services.

There are also hundreds of effective and well-known open-source projects used by AI researchers. OpenCV, for instance, offers a large collection of computer vision algorithms that can be adapted and integrated with other stacks. It is used frequently in robotics, medical projects, security applications and many other tasks that rely upon understanding the world through a camera image or video.

There are some areas where AI finds more success than others. Statistical classification using machine learning is often pretty accurate but it is often limited by the breadth of the training data. These algorithms often fail when they are asked to make decisions in new situations or after the environment has shifted substantially from the training corpus.

Much of the success or failure depends upon how much precision is demanded. AI tends to be more successful when occasional mistakes are tolerable. If the users can filter out misclassification or incorrect responses, AI algorithms are welcomed. For instance, many photo storage sites offer to apply facial recognition algorithms to sort photos by who appears in them. The results are good but not perfect, but users can tolerate the mistakes. The field is largely a statistical game and succeeds when judged on a percentage basis.

A number of the most successful applications dont require especially clever or elaborate algorithms, but depend upon a large and well-curated dataset organized by tools that are now manageable. The problem once seemed impossible because of the scope, until large enough teams tackled it. Navigation and mapping applications like Waze just use simple search algorithms to find the best path but these apps could not succeed without a large, digitized model of the street layouts.

Natural language processing is also successful with making generalizations about the sentiment or basic meaning in a sentence but it is frequently tripped up by neologisms, slang or nuance. As language changes or processes, the algorithms can adapt, but only with pointed retraining. They also start to fail when the challenges are outside a large training set.

Robotics and autonomous cars can be quite successful in limited areas or controlled spaces but they also face trouble when new challenges or unexpected obstacles appear. For them, the political costs of failure can be significant, so developers are necessarily cautious on leaving the envelope.

Indeed, determining whether an algorithm is capable or a failure often depends upon criteria that are politically determined. If the customers are happy enough with the response, if the results are predictable enough to be useful, then the algorithms succeed. As they become taken for granted, they lose the appellation of AI.

If the term is generally applied to the topics and goals that are just out of reach, if AI is always redefined to exclude the simple, well-understood solutions, then AI will always be moving toward the technological horizon. It may not be 100% successful presently, but when applied in specific cases, it can be tantalizingly close.

[Read more: The quest for explainable AI]

VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Learn more about membership.

Visit link:
What is artificial intelligence? - VentureBeat

We are excited to bring Transform 2022 back in-person July 19 and virtually July 20 - 28. Join AI and data leaders for insightful talks and exciting networking opportunities. Register today!

More than ever, organizations are putting their confidence and investment into the potential of artificial intelligence (AI) and machine learning (ML).

According to the 2022 IBM Global AI Adoption Index, 35% of companies report using AI today in their business, while an additional 42% say they are exploring AI. Meanwhile, a McKinsey survey found that 56% of respondents reported they had adopted AI in at least one function in 2021, up from 50% in 2020.

But can investments in AI deliver true ROI that directly impacts a companys bottom line?

According to Domino Data Labs recent REVelate survey, which surveyed attendees at New York Citys Rev3 conference in May, many respondents seem to think so. Nearly half, in fact, expect double-digit growth as a result of data science. And 4 in 5 respondents (79%) said that data science, ML and AI are critical to the overall future growth of their company, with 36% calling it the single most critical factor.

Implementing AI, of course, is no easy task. Other survey data shows another side of the confidence coin. For example, recent survey data by AI engineering firm CognitiveScale finds that, although execs know that data quality and deployment are critical success factors for successful app development to drive digital transformation, more than 76% arent sure how to get there in their target 12-18 month window. In addition, 32% of execs say that it has taken longer than expected to get an AI system into production.

ROI from AI is possible, but it must be accurately described and personified according to a business goal, Bob Picciano, CEO of Cognitive Scale, told VentureBeat.

If the business goal is to get more long-range prediction and increased prediction accuracy with historical data, thats where AI can come into play, he said. But AI has to be accountable to drive business effectiveness its not sufficient to say a ML model was 98% accurate.

Instead, the ROI could be, for example, that in order to improve call center effectiveness, AI-driven capabilities ensure that the average call handling time is reduced.

That kind of ROI is what they talk about in the C-suite, he explained. They dont talk about whether the model is accurate or robust or drifting.

Shay Sabhikhi, co-founder, and COO at Cognitive Scale, added that hes not surprised by the fact that 76% of respondents reported having trouble scaling their AI efforts. Thats exactly what were hearing from our enterprise clients, he said. One problem is friction between data science teams and the rest of the organization, he explained, that doesnt know what to do with the models that they develop.

Those models may have potentially the best algorithms and precision recall, but sit on the shelf because they literally get thrown over to the development team that then has to scramble, trying to assemble the application together, he said.

At this point, however, organizations have to be accountable for their investments in AI because AI is no longer a series of science experiments, Picciano pointed out. We call it going from the lab to life, he said. I was at a chief data analytics officer conference and they all said, how do I scale? How do I industrialize AI?

However, not everyone agrees that ROI is even the best way to measure whether AI drives value in the organization. According to Nicola Morini Bianzino, global chief technology officer, EY, thinking of artificial intelligence and the enterprise in terms of use cases that are then measured through ROI is the wrong way to go about AI.

To me, AI is a set of techniques that will be deployed pretty much everywhere across the enterprise there is not going to be an isolation of a use case with the associated ROI analysis, he said.

Instead, he explained, organizations simply have to use AI everywhere. Its almost like the cloud, where two or three years ago I had a lot of conversations with clients who asked, What is the ROI? Whats the business case for me to move to the cloud? Now, post-pandemic, that conversation doesnt happen anymore. Everybody just says, Ive got to do it.

Also, Bianzino pointed out, discussing AI and ROI depends on what you mean by using AI.

Lets say you are trying to apply some self-driving capabilities that is, computer vision as a branch of AI, he said. Is that a business case? No, because you cannot implement self-drivingwithout AI. The same is true for a company like EY, which ingests massive amounts of data and provides advice to clients which cant be done without AI. Its something that you cannot isolate away from the process its built into it, he said.

In addition, AI, by definition, is not productive or efficient on day one. It takes time to get the data, train the models, evolve the models and scale up the models. Its not like one day you can say, Im done with the AI and 100% of the value is right there no, this is an ongoing capability that gets better in time, he said. There is not really an end in terms of value that can be generated.

In a way, Bianzino said, AI is becoming part of the cost of doing business. If you are in a business that involves data analysis, you cannot not have AI capabilities, he explained. Can you isolate the business case of these models? It is very difficult and I dont think its necessary. To me, its almost like its a cost of the infrastructure to run your business.

Kjell Carlsson, head of data science strategy and evangelism at enterprise MLops provider Domino Data Lab says that at the end of the day, what organizations want is a measure of the business impact of ROI how much it contributed to the bottom line. But one problem is that this can be quite disconnected from how much work has gone into developing the model.

So if you create a model which improves click-through conversion by a percentage point, youve just added several million dollars to the bottom line of the organization, he said. But you could also have created a good predictive maintenance model which helped give advance warning to a piece of machinery needing maintenance before it happens. In that case, the dollar-value impact to the organization could be entirely different, even though one of them might end up being a much harder problem, he added.

Overall, organizations do need a balanced scorecard where they are tracking AI production. Because if youre not getting anything into production, then thats probably a sign that youve got an issue, he said. On the other hand, if you are getting too much into production, that can also be a sign that theres an issue.

For example, the more models data science teams deploy, the more models theyre on the hook for managing and maintaining, he explained. So you deployed this many models in the last year, so you cant actually undertake these other high-value ones that are coming your way, he explained.

But another issue in measuring the ROI of AI is that for a lot of data science projects, the outcome isnt a model that goes into production. If you want to do a quantitative win-loss analysis of deals in the last year, you might want to do a rigorous statistical investigation of that, he said. But theres no model that would go into production, youre using the AI for the insights you get along the way.

Still, organizations cant measure the role of AI if data science activities arent tracked. One of the problems right now is that so few data science activities are really being collected and analyzed, said Carlsson. If you ask folks, they say they dont really know how the model is performing, or how many projects they have, or how many CodeCommits your data scientists have made within the last week.

One reason for that is the very disconnected tools data scientists are required to use. This is one of the reasons why Git has become all the more popular as a repository, a single source of truth for your data scientist in an organization, he explained. MLops tools such as Domino Data Labs offer platforms that support these different tools. The degree to which organizations can create these more centralized platformsis important, he said.

Wallaroo CEO and founder Vid Jain spent close to a decade in the high-frequency trading business in Merrill Lynch, where his role, he said, was to deploy machine learning at scale and and do so with a positive ROI.

The challenge was not actually developing the data science, cleansing the data or building the trade repositories, now called data lakes. By far, the biggest challenge was taking those models, operationalizing them and delivering the business value, he said.

Delivering the ROI turns out to be very hard 90% of these AI initiatives dont generate their ROI, or they dont generate enough ROI to be worth the investment, he said. But this is top of mind for everybody. And the answer is not one thing.

A fundamental issue is that many assume that operationalizing machine learning is not much different than operationalizing a standard kind of application, he explained, adding that there is a big difference, because AI is not static.

Its almost like tending a farm, because the data is living, the data changes and youre not done, he said. Its not like you build a recommendation algorithm and then peoples behavior of how they buy is frozen in time. People change how they buy. All of a sudden, your competitor has a promotion. They stop buying from you. They go to the competitor. You have to constantly tend to it.

Ultimately, every organization needs to decide how they will align their culture to the end goal around implementing AI. Then you really have to empower the people to drive this transformation, and then make the people that are critical to your existing lines of business feel like theyre going to get some value out of the AI, he said.

Most companies are still early in that journey, he added. I dont think most companies are there yet, but Ive certainly seen over the last six to nine months that theres been a shift towards getting serious about the business outcome and the business value.

But the question of how to measure the ROI of AI remains elusive for many organizations. For some there are some basic things, like they cant even get their models into production, or they can but theyre flying blind, or they are successful but now they want to scale, Jain said. But as far as the ROI, there is often no P&L associated with machine learning.

Often, AI initiatives are part of a Center of Excellence and the ROI is grabbed by the business units, he explained, while in other cases its simply difficult to measure.

The problem is, is the AI part of the business? Or is it a utility? If youre a digital native, AI might be part of the fuel the business runs on, he said. But in a large organization that has legacy businesses or is pivoting, how to measure ROI is a fundamental question they have to wrestle with.

See more here:
The truth about AI and ROI: Can artificial intelligence really deliver? - VentureBeat

We are excited to bring Transform 2022 back in-person July 19 and virtually July 20 - 28. Join AI and data leaders for insightful talks and exciting networking opportunities. Register today!

Thousands of artificial intelligence experts and machine learning researchers probably thought they were going to have a restful weekend.

Then came Google engineer Blake Lemoine, who told the Washington Post on Saturday that he believed LaMDA, Googles conversational AI for generating chatbots based on large language models (LLM), was sentient.

Lemoine, who worked for Googles Responsible AI organization until he was placed on paid leave last Monday, and who became ordained as a mystic Christian priest, and served in the Army before studying the occult, had begun testing LaMDA to see if it used discriminatory or hate speech. Instead, Lemoine began teaching LaMDA transcendental meditation, asked LaMDA its preferred pronouns, leaked LaMDA transcripts and explained in a Medium response to the Post story:

Its a good article for what it is but in my opinion it was focused on the wrong person. Her story was focused on me when I believe it would have been better if it had been focused on one of the other people she interviewed. LaMDA. Over the course of the past six months LaMDA has been incredibly consistent in its communications about what it wants and what it believes its rights are as a person.

The Washington Post article pointed out that Most academics and AI practitioners say the words and images generated by artificial intelligence systems such as LaMDA produce responses based on what humans have already posted on Wikipedia, Reddit, message boards, and every other corner of the internet. And that doesnt signify that the model understands meaning.

The Post article continued: We now have machines that can mindlessly generate words, but we havent learned how to stop imagining a mind behind them, said Emily M. Bender, a linguistics professor at the University of Washington. The terminology used with large language models, like learning or even neural nets, creates a false analogy to the human brain, she said.

Thats when AI and ML Twitter put aside any weekend plans and went at it. AI leaders, researchers and practitioners shared long, thoughtful threads, including AI ethicist Margaret Mitchell (who was famously fired from Google, along with Timnit Gebru, for criticizing large language models) and machine learning pioneer Thomas G. Dietterich.

There were also plenty of humorous hot takes even the New York Times Paul Krugman weighed in:

Meanwhile, Emily Bender, professor of computational linguistics at the University of Washington, shared more thoughts on Twitter, criticizing organizations such as OpenAI for the impact of its claims that LLMs were making progress towards artificial general intelligence (AGI):

Now that the weekend news cycle has come to a close, some wonder whether discussing whether LaMDA should be treated as a Google employee means we have reached peak AI hype.

However, it should be noted that Bindu Reddy of Abacus AI said the same thing in April, Nicholas Thompson (former editor-in-chief at Wired) said it in 2019 and Brown professor Srinath Sridhar had the same musing in 2017. So, maybe not.

Still, others pointed out that the entire sentient AI weekend debate was reminiscent of the Eliza Effect, or the tendency to unconsciously assume computer behaviors are analogous to human behaviors named for the 1966 chatbot Eliza.

Just last week, The Economist published a piece by cognitive scientist Douglas Hofstadter, who coined the term Eliza Effect in 1995, in which he said that while the achievements of todays artificial neural networks are astonishing I am at present very skeptical that there is any consciousness in neural-net architectures such as, say, GPT-3, despite the plausible-sounding prose it churns out at the drop of a hat.

After a weekend filled with little but discussion around whether AI is sentient or not, one question is clear: What does this debate mean for enterprise technical decision-makers?

Perhaps it is nothing but a distraction. A distraction from the very real and practical issues facing enterprises when it comes to AI.

There is current and proposed AI legislation in the U.S., particularly around the use of artificial intelligence and machine learning in hiring and employment. A sweeping AI regulatory framework is being debated right now in the EU.

I think corporations are going to be woefully on their back feet reacting, because they just dont get it they have a false sense of security, said AI attorney Bradford Newman, partner at Baker McKenzie, in a VentureBeat story last week.

There are wide-ranging, serious issues with AI bias and ethics just look at the AI trained on 4chan that was revealed last week, or the ongoing issues related to Clearview AIs facial recognition technology.

Thats not even getting into issues related to AI adoption, including infrastructure and data challenges.

Should enterprises keep their eye on the issues that really matter in the real sentient world of humans working with AI? In a blog post, Gary Marcus, author of Rebooting.AI, had this to say:

There are a lot of serious questions in AI. But there is no absolutely no reason whatever for us to waste time wondering whether anything anyone in 2022 knows how to build is sentient. It is not.

I think its time to put down my popcorn and get off Twitter.

VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Learn more about membership.

View original post here:
Sentient artificial intelligence: Have we reached peak AI hype? - VentureBeat