March 23, 2023

Head's up: Sen. Mark Kelly (D-Ariz.), Rep. Mike Gallagher (R-Wis.), and White House Office of Science and Technology Policy director Arati Prabhakar join the lineup for Axios’ second annual What's Next Summit this Wednesday, March 29, in Washington, D.C.! Watch live here.

📣 Situational awareness: IT giant Accenture said it will cut 19,000 jobs, or about 2.5% of its workforce, Reuters reports.

Today's Login is 1,270 words, a 5-minute read.

1 big thing: Where today's generative AI shines

Illustration: Shoshana Gordon/Axios

There's no question that today's generative AI tools are impressive, but it's also clear that the technology is much better at some tasks than others.

Why it matters: Generative AI is being pitched as the solution for everything from boosting productivity to replacing an aging population of workers to making search results more effective. So far, though, it's best when there's less on the line.

Today's generative AI isn't 100% accurate — and it's prone to confidently asserting fiction as fact. So in many ways it's better suited for tasks where there isn't one right answer than for many of the fact-based uses where it's being applied, like search.

  • As tech analyst Benedict Evans put it in a tweet earlier this week: "A 20% error rate in a text generator is 'this thing is telling lies and making stuff up' — but a 20% error rate in an image generator is 'I'm not quite sure where the door is, but honestly who cares when it looks like that.'"
  • Former Windows boss Steven Sinofsky put it more bluntly in a reply to Evans. "The more the focus is on creativity for generative technology the better," Sinofsky said. "Until sourcing and verification are addressed, using it for fact-based works is like using AutoCorrect without backspace/undo in that the errors will just pile up."

That said, here are three areas where the technology is already highly useful:

Creative tasks

Being able to describe a scene and have it come to life will fundamentally shift how movies, video games and other works are created. Programs like Dall-E 2, Stable Diffusion and Adobe's new Firefly are already quite good at creating powerful images from a few words of text.

  • Generative AI doesn't have to deliver a finished product to be useful. Artists and writers who have embraced the technology praise its ability to spark new ideas and serve as a launchpad for human creativity.
  • Of course, plenty of legal and ethical issues remain, from copyright issues over how the engines were trained to the propensity for the output to perpetuate bias and stereotypes included in training data.

Genre shifting

Today's text generators, engines like OpenAI's GPT-4, are already quite good at taking a set of data and presenting it in other ways, from summarizing meeting notes to transforming information into a fact sheet, press release or a series of tweets.

  • On the fun side, ChatGPT and Google's Bard can take information and serve it up as poems, raps or virtually any other form of expression.

Experimenting

Even if you're wary of the technology because of its limits and drawbacks, it's worth poking around at it now, before it's fully baked, just to understand its capabilities.

  • Companies say that's part of why they've given the technology to the public even in its highly imperfect state — so they can see how people use it, and so all of society can better understand what's headed our way.
  • It's a chance for people to register feedback and civil rights groups to point out flaws in the technology.
  • It's also a chance for regulators to catch up quickly, something critics say is urgently needed. As we've written, there is little AI-specific legislation on the books right now.

What's next: Generative AI is rapidly improving and may soon be able to handle tasks it can't reliably do today.

  • It will someday soon be able to integrate tightly with a company's own data and intellectual property. Salesforce, Microsoft, Google and others are already pitching today's technology for some of these uses and prediciting far more ambitious future applications.
  • For the most part, companies are using the tools to "draft" text and code, giving humans a chance to modify the suggestions before anything is sent or committed. Microsoft describes its workplace tools as "copilots," while others are using a similar approach.

2. Google's Bard lists chatbots' strong suits

Screenshot: Axios

I asked Bard, Google's new chatbot, what it thinks generative AI is good for today.

  • For better, or worse, it largely agreed with my analysis.

Bard's "thought" bubble: Here's what the chatbot had to say, taken verbatim:

Generative AI is good at tasks that require creativity, such as:
— Creating art and music
— Writing stories and poetry
— Designing products and services
— Generating new ideas
Generative AI is also good at tasks that require knowledge of a particular domain, such as:
— Translating languages
— Answering questions
— Creating code
Generative AI is still under development, but it has the potential to revolutionize many industries. For example, it could be used to create personalized content for each user, or to automate tasks that are currently done by humans.

Yes, but: Bard isn't always a good judge, even of itself. Already this week it has told people, for example, that the engine was trained by reading people's Gmail — something the humans at Google assure me is not the case.

3. FTC wants to make canceling subscriptions easier

Illustration: Shoshana Gordon/Axios

The Federal Trade Commission is proposing amending a federal rule so consumers can more easily cancel recurring subscriptions, Axios' Ashley Gold reports.

Why it matters: The proposal, announced Thursday, aims to cover "everything from cosmetics to newspapers to gym memberships" and is another move by FTC chair Lina Khan to try to protect consumers in a digital economy.

  • The agency wants a "click to cancel" provision, requiring sellers to make cancelling as easy as signing up for a service.
  • Stricter rules from the FTC around subscriptions may draw the ire of the business community, which is already battling with the agency over what they think are unfair and overreaching policies.

What they're saying: "Companies should not be able to manipulate consumers into paying for subscriptions they don't want," Khan said on a call with reporters ahead of the release of the proposed rule. "This is a semi-universal experience for American consumers, and we have the legal authority to do this."

  • "We've seen companies deploying dark patterns and manipulative design tactics ... we want to make sure companies aren't tricking consumers into signing up in the first place," she said. "It should be as easy to cancel as it is to sign up."

How it works: The FTC proposes amending the Negative Option Rule, established in 1973 to fight against unfair or deceptive practices around subscriptions or memberships.

  • Violations of the proposed rule would carry a $50,000 fine per day of noncompliance, Khan said.

Details: The FTC says it receives thousands of consumer complaints yearly about practices such as making people cancel services in person or keeping them on hold with customer service before cancelling.

What to watch: Khan said a dramatic growth in subscription-based business models, like the proliferation of streaming services, increases the urgency for the FTC to make such a move.

  • The agency will receive public comments on the proposal before anything becomes final.

4. Take note

On Tap

  • TikTok CEO Shou Zi Chew will testify before the U.S. House Committee on Energy and Commerce beginning at 10am ET. It follows the Biden administration’s urging that China-based owner ByteDance sell TikTok to avoid an all-out ban.

Trading Places

  • The Information reports that Meta has rehired Nick Grudin, a veteran manager who had left Facebook's parent company last year for an executive role at NFT startup Dapper Labs.

ICYMI

  • Microsoft launched a public preview of Loop, its would-be rival to Notion and Asana. (TechCrunch)
  • This year's Turing Award — the highest prize in computer science — was given to Bob Metcalfe, the co-inventor of Ethernet. (NY Times)

5. After you Login

  • And since we are on an animal kick this week, we'd be remiss not to pass along our congratulations to Mr. Pickles, the endangered Houston Zoo tortoise who just became a father at age 90.

Thanks to Scott Rosenberg and Peter Allen Clark for editing and Bryan McBournie for copy editing this newsletter.