Good morning. I spoke with Sol Rashidi, an AI expert who helped IBM launch Watson back in 2011, about strategies and challenges for AI deployment in the enterprise.

She said that the field has changed a lot in the past 10 years; the biggest change, of course, was the launch of ChatGPT, which made generative AI accessible in a way it’s never been before.

Read on for the full story.

— Ian Krietzberg, Editor-in-Chief, The Deep View

In today’s newsletter:

• 🌎 AI for Good: The United Nations’ World Environment Situation Room

• 📊GenAI adoption slows amid enterprise concerns over inaccurate responses

• 🛜 Vimeo, Etsy and the AI wars

• 🏢In conversation with Sol Rashidi: The human problem with AI deployment

AI for Good: The United Nations’ World Environment Situation Room

In 2022, the United Nations Environmental Program (UNEP) launched its World Environment Situation Room (WESR), a digital platform that leverages AI to collate and organize climate-related data. 

The details: This system enables both (near) real-time analysis and future predictions of sea level rise, changes in glacier mass and the concentration of carbon dioxide in the atmosphere, among other things. 

• Within this system is the International Methane Emissions Observatory, which uses AI to integrate and track global methane emissions from a variety of sources. 

• UNEP said that it hopes for its WESR to become “a mission control center for planet earth, where all of our vital environmental indicators can be seamlessly monitored to drive actions.”

The organization added that AI tools can be used to track the environmental footprints of specific products and their supply chains. 

Why it matters: “Reducing the energy sector’s methane emissions is one of the quickest, most feasible and cost-effective ways to limit the impacts of climate warming, and reliable data-driven action will play a big role in achieving these reductions,” David Jensen, coordinator of the UNEP’s digital transformation’s taskforce, said.

Build custom AI chatbots effortlessly

Chatling enables you to build and deploy advanced AI chatbots on any website in less than 15 minutes, using a simple drag-and-drop interface.

With Chatling’s no-code AI platform, you can:

• Automate tasks like customer support, onboarding, lead generation, and sales

• Train your chatbot on a website, knowledge base, and documents

• Engage with your global audience in over 85 languages

👉Create your AI chatbot now for free.

GenAI adoption slows amid enterprise concerns over inaccurate responses 

Cost and accuracy, according to a new Lucidworks study, are among the more significant barriers to generative AI adoption in the enterprise. 

The details: Lucidworks’ second annual Generative AI Global Benchmark Study found that 44% of manufacturers are increasingly worried about the accuracy of GenAI output, something Lucidworks called a “legitimate concern with potentially disastrous consequences.” 

• “While many manufacturers see the potential benefits of generative AI, challenges such as response accuracy and cost are causing them to take a more cautious approach,” Lucidwork’s CEO Mike Sinoway said. 

• “This is reflected in spending plans, with significantly fewer (60%, compared to 93% last year) planning to increase AI investments compared to last year.” 

Why it matters: This — cost and accuracy — is the fundamental challenge of the business of AI. VCs are betting that it will be solved, and in glorious fashion, as funding to AI startups spiked heavily in the second quarter of the year. 

But an increasing number of analysts and investors are concerned about those two points — cost and accuracy — which could be the two-pronged maneuver that brings the hype-inflated industry crashing back to earth. 

• Join us for The AI Conference 2024

  • Share 2 Days with the Brightest Minds in AI including scientists and engineers from OpenAI, Meta, Deepmind, and many more. This week only — Register using discount code: deepview24 for a $350 discount!*

• Effortlessly integrate customer-specific data from dozens of apps, like AI21 and Writesonic do, with 70% less engineering. Try Paragon for free or get a demo today!*

• This cement startup says it goes beyond net zero to true zero carbon emissions (CNBC).

• Elon Musk Says Second Neuralink Brain Implant About a Week Away (Bloomberg).

• How Amazon Built a Key AI Team (The Information).

• Global EV sales up 13% in June, down 7% in Europe, Rho Motion says (Reuters).

• The most important sign that US inflation may stay down for good (Semafor).

Vimeo, Etsy and the AI wars

Earlier this week, Etsy updated its policies to account for generative AI, saying in a blog post that sellers will be allowed to sell AI-generated art, so long as they disclose that it was AI-generated. 

• But while sellers are now allowed to sell AI-generated content, they can’t sell “prompt bundles” on the platform. 

• The prompts, Etsy said, are an “integral part of the creative process”; selling them would “undermine the value of the artist's creative input and curation.” 

This move comes as the major question of whether it is a legal violation of existing copyright law to train AI models on art, photos and text without artists’ knowledge, permission or compensation remains unanswered (and in litigation); Etsy acknowledged this, saying it will be “closely monitoring developments” in this arena. 

• “Etsy has made it clear they don't care about artists,” professional illustrator Reid Southen said in response.

Now, just a few weeks ago, video-sharing platform Vimeo — also in a blog post — published its stance on AI, saying that it won’t allow AI models to be trained using the videos hosted on its platform without users’ “explicit consent.” 

Vimeo added that it prohibits unauthorized content scraping by model developers and implements “security protocols designed to protect user-generated content.” 

Why it matters: There are lines being drawn — as we are about to get into below — between the companies who have decided to fully embrace AI, period, and the companies who have deemed it a smarter strategy to build user trust by not embracing AI. 

This tale of Vimeo and Etsy exemplifies those different approaches. 

Invest smarter with Public and AI

If you're looking to leverage AI in your investment strategy, you need to check out Public. The all-in-one investing platform allows you to build a portfolio of stocks, options, bonds, crypto, and more, while incorporating the latest AI technology to help you make more informed investment decisions.

Last year, Public introduced Alpha, a proprietary AI that transforms how you interact with your portfolio. Alpha delivers real-time financial insights on the assets you care about, including alerts on unusual market movements and earnings call summaries. This proactive approach ensures you're always informed about significant events affecting your investments.

You can also swipe down in the Public app to access Alpha's reactive chat feature. Ask any question about any asset, and Alpha leverages its natural language processing capabilities to provide insightful answers with precise fundamental data.

Join Public, and build your primary portfolio with AI-powered insights and analysis.

In conversation with Sol Rashidi: The human problem with AI deployment 

In the enterprise, generative AI is beginning to move out of the experimentation phase and into the real-world deployment phase. Deloitte’s second-quarter State of Generative AI in the Enterprise report — which surveyed nearly 2,000 executives around the world — found that businesses are focused now on deriving actual, tangible value from their genAI experiments.

• Still, only around a third of overall organizations are seeing these benefits (in increasing revenue, boosting efficiency, cutting costs, improving products and increasing the pace of development). 

But what the report found is that organizations with “very high” levels of expertise in AI are moving much more aggressively on AI. And much higher percentages of these organizations are achieving the kind of tangible benefits their leaders have deemed so important. 

The core of the divide between the companies having success with AI and those that are not has to do with corporate strategies and the humans that create them, according to Sol Rashidi, an AI expert who has achieved numerous accolades in the space, including inclusion in 2022’s “50 Most Powerful Women in Tech.” 

I sat down with Rashidi — who helped IBM launch Watson in 2011 — to discuss the challenges of deployment in today’s environment. 

How AI deployment has changed in the past 10 years: Artificial intelligence is not a new technology. It got its start back in the ‘50s and experienced a surge of popularity in the ‘70s (which is when the term “AI” was coined). IBM’s launch of Watson in 2011, according to Rashidi, marked the “first time” that these academic concepts entered the corporate world. 

• The difference between now and then, she said, is that in 2011, AI tools and applications were almost exclusively business-to-business. The software, in part due to its enormous price tag, was “not democratized to small companies, midsize companies or consumers.”

The release of ChatGPT at the end of 2022 made it accessible to everyone. 

“I think the biggest pivot is it's no longer focused on enterprise, but now everyone can leverage it,” she said. 

Deployment is a human problem, not a technical one: And while ChatGPT might have solved the problem of accessibility, the means of adoption come down to human strategy — and not all strategies should be made alike when it comes to AI. 

• “People forget that you have to connect your strategy to the maturity of your organization,” Rashidi said. “If you happen to be a company that’s very risk averse, your ability to deploy AI-based capabilities is also very limited.” 

She said that a common mistake companies make is to forcibly attempt to apply AI to create business value when it can instead be applied in targeted internal ways. Using AI to save a given organization money by making internal processes less costly, according to Rashidi, creates a “tremendous amount of business value.” 

It’s not all about outward-facing deployment.  

Sometimes the best AI strategy is no AI strategy: And in the midst of this feverish push toward more and more AI, there has been a rising backlash against it. Artists don’t want their content to be used to train AI systems, and many consumers are disinterested in using systems that are untrustworthy and biased. 

While many organizations are pushing harder on the genAI hype train, this surge has led others to go the opposite direction; Dove in April pledged to “never use AI to represent real women in its advertising.” 

• Cara, an Instagram alternative that promises to protect artists from AI scraping, surged from 40,000 users to around one million in just a few days. Not By AI offers a badge that asserts that a piece of content was made by humans, not machines; it claims 264,000 websites are currently using it.  

Rashidi said that sometimes, the best AI strategy is no AI strategy. 

“Why use a sledgehammer when a hammer does the trick? AI can be overkill and sometimes AI is not the answer,” she said. “That's just the simple reality of it.”

A poll before you go

Thanks for reading today’s edition of The Deep View!

We’ll see you in the next one.

Your view on whether AI progress should slow down:

Around 35% of you agreed with the idea that a cautious approach is preferable to being the first country to develop more powerful AI systems.

And around 30% of you said the opposite.

Agree:

• “That's (the) general rule of science. Science should have no place for people without patience, diligence, caution and ability to accept mistakes and rollback. (The) story inside the lab is always different from (the) story outside the lab. Corporations should refrain from putting target pressure on engineers and researchers.”

Disagree:

• “I don't believe our Congress has the capacity to create or implement appropriate rules to regulate the development. The attempt would degenerate into either a political debate or a ‘follow the money’ sellout — either of which would create more problems than it would solve.”

_{*Public disclosure: All investing involves the risk of loss, including loss of principal. Brokerage services for US-listed, registered securities, options and bonds in a self-directed account are offered by Public Investing, Inc., member FINRA & SIPC. Cryptocurrency trading services are offered by Bakkt Crypto Solutions, LLC (NMLS ID 1828849), which is licensed to engage in virtual currency business activity by the NYSDFS. Cryptocurrency is highly speculative, involves a high degree of risk, and has the potential for loss of the entire amount of an investment. Cryptocurrency holdings are not protected by the FDIC or SIPC.}

_{Alpha is an experiment brought to you by Public Holdings, Inc. (“Public”). Alpha is an AI research tool powered by GPT-4, a generative large language model. Alpha is experimental technology and may give inaccurate or inappropriate responses. Output from Alpha should not be construed as investment research or recommendations, and should not serve as the basis for any investment decision. All Alpha output is provided “as is.” Public makes no representations or warranties with respect to the accuracy, completeness, quality, timeliness, or any other characteristic of such output.  Your use of Alpha output is at your sole risk.  Please independently evaluate and verify the accuracy of any such output for your own use case.*}