Innovating with AI: TikTok's New AI Chatbot Tako & Google's Generative AI in Search

Today:

TikTok is testing an in-app AI chatbot called ‘Tako’

TikTok is giving the AI chatbot biz a whirl with its new experiment, Tako. Limited test? Check. Select markets? Check. On the right side of your screen, ready for a friendly chit-chat? You betcha. Tap Tako and ask away about any video that's left you scratching your head or if you need new content to binge on.

Word on the street is Tako was found in action by the smart cookies over at Watchful.ai. TikTok has confirmed that, yep, they're testing Tako, but it's not quite ready for the big time in the U.S. It's got its training wheels on in other global markets, though, and it won't be popping up for the kiddos on the app.

Just like other companies, TikTok lets users delete their chats to alleviate those privacy woes. Still, it's murky if Tako is connecting the dots between your chats and your personal info, or what its long-term data retention policies are.

Long story short, TikTok's Tako is still fresh out of the oven, folks. Stay tuned for more updates on this chatty newcomer.

READ THE ARTICLE ON TECHCRUNCH.

Google starts opening up access to its new generative AI capabilities in Search

Seems like Google's playing show and tell with its shiny new AI tech for Search. Remember that teaser from the I/O event couple weeks back? Well, they're starting to let people play with the new toys in their sandbox.

It's all about this thing called Search Labs. Fancy name for a new program where you can sign up, experiment with their new stuff, like the 'Search Generative Experience,' before the masses get their hands on it.

You gotta sign up to get in on this deal. Once you're in, they're gonna let you test drive these new features they've been cooking up. Basically, it's all about making your life a tad easier when it comes to digging up info.

Like, let's say you're thinking about picking up the ukulele or guitar. Normally, you'd have to do a bunch of separate searches to figure out which one's right for you. But this AI-powered Search is smart enough to pull it all together. It'll give you the low-down in one shot.

This ain't just for big stuff though. Even for the small potatoes, like getting out that stubborn coffee stain on your favorite wool sweater or hustling through a passport renewal. This thing can toss out quick tips, show you your options, and help you take the next step.

They're even trying to help out with shopping. Like, say you want to redo your kitchen with some peel and stick wallpaper. Search can fill you in on the important stuff, like how hard it'll be to get that stuff off your walls later, while throwing you a list of options to pick from, with prices, customer ratings, and even purchase links.

If you're still curious, you can keep the chat going by asking follow-up questions. The AI keeps the context in mind, making it feel like an actual conversation.

Just remember folks, this is all still in the works. It's like Google's first baby step in integrating this fancy generative AI into Search. They're itching to see what we think of it and make it even more user-friendly. So, let's go give it a whirl, shall we?

READ THE ARTICLE ON TECHCRUNCH.

Nvidia races towards $1tn club as AI frenzy drives chip stocks higher

Nvidia, the Silicon Valley's hotshot chipmaker, has hit a streak of luck so good that it's closing in on the $1 trillion club. They're riding high on the wave of folks clamoring for their artificial intelligence processors, and their stock soared by a whopping 27% after they revealed a sales forecast that left Wall Street estimates eating dust.

They managed to add nearly $200 billion to their market value after their latest report, which is more than the total worth of Intel or Qualcomm. That's like having your piggy bank suddenly weigh as much as an elephant. They're now just a hop, skip and a jump away from rubbing shoulders with the likes of Apple, Microsoft, Alphabet, Amazon and Saudi Aramco in the uber-rich group.

Nvidia's success seems to be contagious, with chip suppliers TSMC and ASML riding up the ranks too. They've been bragging about their tech's ability to meet the massive demand for building AI systems that can cook up human-like content. Industries from cars to healthcare to telecoms are tripping over each other to grab their power-hungry chips.

Their top-dog H100 processor is flying off the shelves, wanted by everyone from big-league tech companies to fresh-faced AI start-ups. "Nvidia is at the front line of the AI demand surge," said Geoff Blaber, big boss at CCS Insight. He likened their chips to the "picks and shovels" of a "massive shift in AI".

Other chipmakers enjoyed a good day too, with AMD, Micron, Microsoft, and Google seeing their stocks shoot up. Even a few Japanese chip equipment suppliers got in on the fun.

However, not all is sunshine and rainbows. Intel, seen as the slowpoke in the AI race, saw its stock dip by 5% as investors are betting on AI shaking up the whole datacentre tech game.

Nvidia's stock has seen a roller coaster ride over the years, linked to the rise and fall of trends like cryptocurrencies and self-driving cars that didn't live up to the hype. But Nvidia's CEO, Jensen Huang, feels like they've hit the jackpot with the rising demand for generative AI. It's like they've been training for this race for 15 years and now they're sprinting ahead, leaving others to eat their dust.

READ THE ARTICLE ON FINANCIAL TIMES.

With new grant program, OpenAI aims to crowdsource AI regulation

OpenAI is playing a game of 'Who Wants to Be a Hundred-Thousandaire' by offering up ten grants worth $100,000 apiece. The big idea? Finding a democratic way to decide what sort of AI systems should and shouldn't pull, of course "within the law" — there's no Wild West happening here.

This all comes hot on the heels of OpenAI's plea for an international regulatory body to put AI in check, kinda like what we do with nuclear power. It's like they're saying, "This AI stuff's moving so fast, our current rules can't keep up!"

So now, OpenAI's throwing out cash to those bright enough to cook up a democratic process for AI boundaries. This ain't about making binding decisions just yet. It's about exploring questions and cooking up new-fangled democratic tools. Sort of a "see what sticks" approach to handling super-smart AI.

OpenAI's dream is to create a democracy-style process where a mixed bag of folks can have a good ol' chinwag and then decide what happens next. They're looking to answer brain-busters like "Should AI systems throw shade at public figures?" and "How do we show different viewpoints in AI outputs?"

While this might all look like OpenAI's acting out of pure-hearted concern for the public interest, it's worth noting that this comes just after OpenAI's big cheese, Sam Altman, had a few choice words to say about the EU's proposed AI regulations. Makes you wonder if there's a bit of self-interest simmering beneath this generous gesture.

But hey, self-serving or not, it's an interesting spin on how we manage AI. It'll be a hoot to see what kind of democratic ideas folks come up with and who ends up pocketing the cash. Want in? You've got until June 24 to throw your hat into the ring. If you bag one of the grants, you'll have to pull together a concept with at least 500 folks, share your findings by October 20, and give away your code. So go on, get cracking!

READ THE ARTICLE ON TECHCRUNCH.

How do we best govern AI?

Well, here we have Brad Smith from Microsoft saying, "Hold your horses, folks, let's not just ask what these smarty-pants computers can do, but also ponder over what they should be allowed to do." Why? 'Cause computers are now butting into our territory, calling the shots that were once ours.

So, every Tom, Dick, and Harry is scratching their noggin about how to handle and harness this revved-up tech. Microsoft, not to be left in the dust, is adding their two bits, guided by the ethical AI roadmap that their head honcho, Satya Nadella, started drafting in 2016.

Since then, they've been busier than a one-legged man in a butt-kicking contest, creating rules for AI and even assembling an army of 350 to keep AI from turning into a 21st-century horror show. Their AI's been quite the good Samaritan - saving folks' eyesight, duking it out with cancer, protecting folks from bad weather, and even giving 13-year-olds a leg up with their pesky math homework.

But Brad's no pie-in-the-sky dreamer. He's seen how social media, as harmless as it seemed, turned into a two-faced monster, and he doesn't want a repeat performance with AI.

He's hammering home that we need to draw some lines in the sand for AI, to keep it from going rogue. But, the tech eggheads can't do this solo; everyone needs to chime in.

Accountability - that's the cornerstone of Microsoft's AI ethos. This means making sure machines toe the line with humans, and the humans behind the machines answer to the rest of us. And no matter how brainy AI gets, it always stays on a human-held leash.

In the end, Brad brings up the law. Just like no one's above it, neither should any tech be. The million-dollar question: How do we get AI under this umbrella, and what new rules and regs do we need to whip up for it?

So, here's the skinny. The piece is all about trying to saddle this wild stallion we call AI, without spooking it.

They've cooked up a five-part game plan for this.

• Beefing up safety rules, such as adding extra padding to linebackers, based on a framework proposed by the U.S. National Institute of Standards and Technology.

• Implementing safety measures for AI to ensure key systems like electricity and water are not compromised.

• Establishing a legal structure for AI to address accountability and responsibility, similar to determining pet ownership responsibilities in a neighborhood.

• Promoting openness by involving academics and nonprofits in the development and regulation of AI, allowing broader access to its benefits.

• Encouraging public-private collaboration to address societal challenges associated with emerging technologies, akin to a community meeting to resolve a local issue.

They also spill the beans about their own wrangling with AI at Microsoft. They've got ethical guidelines, corporate rules, a heap of case reviews, and even a program to keep a close eye on the tricky uses of AI. It's been a wild ride, they admit, but it's all about keeping AI in check while letting it strut its stuff.

So, to put it simply, this article is a wake-up call. It's telling us that, while AI is a real whiz-bang tool, we gotta make sure it doesn't morph into a loose cannon causing a ruckus.

READ THE ARTICLE ON MICROSOFT.

OpenAI could ‘cease operating’ in Europe over AI laws: Sam Altman

OpenAI bigwig, Sam Altman, is chewing his nails over the EU's AI laws. He's saying his tech baby might have to close shop in Europe if they can't dance to the EU's tune. He's talked about this after a chinwag at University College London.

The thing is, the EU AI Act, slated to kick in next year, is set to keep AI use in check, especially for high-stakes stuff like picking out job candidates or running medical gear. Altman isn't saying the law's rotten, just that the nitty-gritty of it matters.

Now, the EU's been playing traffic cop with AI for a couple of years, first wheeling out this act in 2021. Since OpenAI's ChatGPT and GPT-4 started making waves, the Euro bigwigs are looking to stretch the law to cover more AI uses.

For instance, chatbots gotta wear a sign saying "I'm AI, not a human," which isn't a thing stateside. They also want to clamp down on face recognition tech and tighten the screws on generative AI, including data transfer rules.

Altman seems to fancy a midway approach between Europe and the US, but Uncle Sam doesn't seem keen on walking the same path. He even gave Congress an earful last week, warning about the potential pitfalls of AI.

In other news, ChatGPT debuted on the Apple App Store last week, and boy, it's been a hit! It's sitting pretty at No. 1 in the "Productivity" category, and No. 3 overall, even beating out TikTok, Instagram, and WhatsApp. But while there's no US law reining in such AI tools, heavyweights like Apple, JPMorgan Chase, and Verizon have told their folks to keep their hands off the language model.

So, we're in a pickle - AI is advancing like a wildfire, but the rulebook's still being written.

READ THE ARTICLE ON NEW YORK POST.

5 ways Windows 11’s new Copilot AI assistant will make your life easier

Microsoft's got a new trick up its sleeve, y'all. They call it "Windows Copilot", and it's some fancy AI assistant for Windows 11.

Imagine a one-stop-shop for all your tech needs, like a trusty hound at your side. It’s like your digital "know-it-all" buddy for your PC and its apps, but without the annoying know-it-all part. It’s Microsoft's answer to OpenAI's ChatGPT, and man, it's like they injected Bing Chat with some high-octane AI fuel.

Copilot’s got some new digs on the right of your screen, where your notifications usually hang out, kinda like that chatty neighbor who won't stop leaning on your fence. And this thing doesn’t just chat, it can interact with your photos and files too.

The bigwigs at Microsoft showed off how this smarty-pants AI can help you with all sorts of problems, from planning a fishing trip to interpreting some code. It's like your own personal geek squad on speed dial.

One neat trick it has is helping you control your Windows environment. Say you're struggling to find how to switch to dark mode, Copilot can do that for you, no sweat. If you’re wondering about reducing eye strain, just ask, and it’ll serve up a solution, pronto.

Windows Copilot can also work across documents and even help you create content. It's like having a virtual assistant who actually knows what they're doing. Want music to work to? Copilot will pop open Spotify faster than you can say "play that funky music".

But here's where things get interesting. Copilot can even help you design a logo for your business using Adobe Express and, get this, send it to a coworker using Teams. It's like having your own creative department at the touch of a button.

And for us folks who like to gab, Copilot can transcribe voice recordings, which is a handy tool, especially if you've got the memory of a goldfish like me.

But all these bells and whistles beg the question: can your PC handle all this AI wizardry? You see, Copilot might demand some serious hardware. It’s like wanting to go off-roading with a sedan.

Microsoft is working with the big boys - AMD, Intel, and Nvidia - to make sure their hardware can keep up with all this AI stuff. They've even got this thing called "Hybrid Loop" that lets your PC use as much of its own juice as possible before asking for a top-up from the cloud.

So, will Copilot work on your PC? That's the million-dollar question. Microsoft's staying mum for now, but I'd bet my last doughnut that they're cooking up a solution. Just be ready, you might need to shell out for a beefier PC.

READ THE ARTICLE ON PC WORLD.

Apple Takes Its 30% Bite From AI Innovation, Costing OpenAI Millions

OpenAI's ChatGPT is the new belle of the ball, wowing folks left and right with its AI smarts. But, there's a twist: Apple, known for not being at the AI vanguard, is still pocketing a pretty penny from it all.

You see, OpenAI launched a shiny new iPhone app, and it's been raining dollars and accolades ever since. Faster than a greased pig in a mud race, it scampered to the top of the App Store, even getting a gold star from Apple itself.

But here's where the apple gets wormy. Apple has a reputation for dipping its beak in the profits, taking a 30% cut of all new subscriptions - the so-called "Apple Tax." The folks at OpenAI didn't seem too fazed, though, accepting the terms and using Apple's own in-app purchasing system.

Do the math and you'll see that Apple nabs a cool $6 out of every $20 ChatGPT Plus subscription. Not bad for a company that hasn't broken a sweat in the AI field.

Despite the rumblings about data privacy, ChatGPT's still available on the App Store, fattening Apple's wallet which, by the way, hit a record $20.8 billion last quarter. However, not everyone's laughing. The feds ruled that Apple's 30% fee violated California's Unfair Competition Law. OpenAI's sticking to their guns, though, likely because Apple's payment system is as smooth as butter.

Things got a tad hairy for Apple when they beat Epic Games in court but now they're flirting with allowing third-party app stores in the EU, thanks to new rules aimed at leveling the playing field.

It's a real barn dance between AI hotshots, tech bigwigs, and the law. Right now, it seems like Apple's calling the tune, piggybacking on others' innovations to line their own pockets, even amidst some heavy eyebrow-raising.

But let's not forget, folks, the AI revolution is only just kicking off, and the mad scramble to turn a profit is only heating up. As for the rest of us? We're just along for the ride.

READ THE ARTICLE ON DECRYPT.

The official ChatGPT app is now available in 11 more countries

OpenAI is kicking down the international doors with its ChatGPT mobile app. First, it was only us Yanks that could have all the fun with this techno-wizardry on our iPhones. Now, they've decided to spread the love to a bunch of other spots around the globe.

This ChatGPT thing is a free app, no pesky ads, thank heavens! For folks who've had a go at it before, it's same old, same old. Just you, the chatbot, and all your deep, dark queries.

So, who are the lucky ducks getting the app? Albania, Croatia, France, Germany, Ireland, Jamaica, New Zealand, Nicaragua, Nigeria, South Korea, the U.K., and us. No Android love yet, but they swear it's on the way.

You use it like you would any messenger app. There's a spot to type in, you hit send, and presto - AI spits back some wisdom. If you're in a lazy or a secret agent mood, you can use your voice instead of typing. And yes, you can keep track of all the profound stuff you've been chatting about, unless you decide to go off the grid.

For those of you who've shelled out twenty bucks a month for ChatGPT Plus, you can tap into GPT-4's brainpower on your phone too. Expect snappier responses, and yes, it's available as an in-app purchase worldwide, albeit at different price tags.

The timing's a hoot with OpenAI's head honcho, Sam Altman, schmoozing with bigwigs across Europe this week. Guess, "Just download the app," is his new calling card. Well played, Sammy boy. Well played.

Remember folks, you saw it here first. If you don't have the app, grab it. If you do, keep asking it weird stuff. It probably won't judge you. Probably.

READ THE ARTICLE ON TECHCRUNCH.