Beyond Tomorrow: OpenAI’s Voice Mode Is Out, Perplexity’s Publishers’ Program — Weekly AI Roundup

Hello there!

We're back with another edition of Beyond Tomorrow. What a week it's been, huh? I'd quote 30 Rock here, but while its not a Wednesday, the sentiment still applies. A bunch of big news came out in the world of artificial intelligence over the course of the last seven days.

Apple's AI features are going to take a little longer than expected, according to our favourite Apple reporter Mark Gurman. The company is really hoping that Apple Intelligence is going to boost the iPhone 16's sales number, reads a Bloomberg report.

Meanwhile, OpenAI is rolling out its Advanced Voice Mode. Its currently in alpha and still only available to a limited number of users, but the company plans to roll out the feature to all ChatGPT Plus users eventually after more testing. For all those of you trying to make Her a reality, please go touch grass.

Are publishers finally going to get paid for their content? Controversial AI startup Perplexity sure is trying, despite its missteps in the past. Seems like an effort made in good faith, but what does that even mean anymore?

Here is some AI news from the week gone by that's worth noting...

OpenAI Begins Rolling Out Advanced Voice Mode

It's official, OpenAI's Advanced Voice Mode is finally coming to users. Or at least to a “small group of ChatGPT Plus users,” according to the company's Twitter.

The company has released the new mode in alpha to some users and has provided instructions on how to access it. Back in May this year, OpenAI had unveiled the new voice mode and said it would be rolled out “in the coming weeks,” according to their blog post on their website.

We're starting to roll out advanced Voice Mode to a small group of ChatGPT Plus users. Advanced Voice Mode offers more natural, real-time conversations, allows you to interrupt anytime, and senses and responds to your emotions. pic.twitter.com/64O94EhhXK

Cut to June, when the company on their official Discord server said that they'd be pushing back the release of the voice mode to iron out some more issues and was pushing the launch back to July. At the time, OpenAI said they were working on improving the model's ability to “detect and refuse certain content.”

True to their word, the Alpha for Advanced Voice Mode began rolling out right at the end of July. The company says that since they first demoed Voice Mode, they've been working to “reinforce the safety and quality of voice conversations”. OpenAI says they've tested GPT-4o's voice capabilities with over 100 external red teamers across 45 languages.

Red teams are generally groups of users acting as threats to test the limits of something. They're usually employed to test for vulnerabilities in software, but in general is industry agnostic. Heck, even as journalists we have red teams to poke holes in our stories so that we can make sure they're watertight.

Back when OpenAI first announced the voice mode for ChatGPT-4o, people collectively lost their minds. The tech seemed promising and the applications endless. The company did, however, become subject to controversy when people pointed out one of the voice models sounded eerily like Scarlett Johansson.

Courtesy: imgflip.com

After a little bit of hullabaloo in the news cycle, Sam Altman announced the removal of the offending voice. The removal was despite both Altman and the voice actress coming out saying that they hadn't used Johansson's voice. By then, the news cycle had moved on with everyone collectively shrugging and saying “okay.”

Perplexity Launches Publisher's Program

Controversial AI startup Perplexity has finally come out with a partnership program to include some of the world's largest publishers. In a blog post released on July 30, the “AI-powered answer engine” announced the Perplexity Publishers' Program.

The project is “designed to promote collective success” according to the post from Perplexity. Publishers included in the company's first batch of the project include: TIME, Der Spiegel, Fortune, Entrepreneur, The Texas Tribune, and WordPress.com.

Perplexity's program includes a revenue sharing agreement with the publishers, which is the most notable in its entire blog post. Right off the bat, the company says that they're adding advertising to their related questions feature.

On the one hand, understandable. AI startups (or any startup) require bucketloads of money. Addressing the monetisation of AI has been a consistent theme for a while now, and there's no right answer. What Perplexity is doing is… complicated.

Pages is a feature from Perplexity that lets users get a response to a query in the form of a formatted article, which cites where it produced the information from.

The way this reads looks like certain responses are more favoured, versus others, as long as you as a publisher or advertiser pays for it. In other words, brands are being allowed to push their views on people who just want answers. The blog doesn't really make clear what Perplexity means.

Has anybody told these companies that their lack of transparency is annoying? (Source:Tenor)

Perplexity says that whenever it makes money from an interaction that includes a publisher's content, the publisher will also make money. Again, the AI company hasn't specified what the revenue share ratio looks like.

In addition, the company is providing publishers access to their Application Programming Interface to integrate the AI-startup into their own technology ecosystems. Perplexity's technology can be leveraged by publishers to create their own custom answering engines on their website, which only cite their content.

Look, it's a step in the right direction, no doubt about that. But has the damage to Perplexity's reputation already been done, with their poor handling of plagiarism in the past? Perhaps. Will that change the fact that publishers are still signing partnerships like this? Unlikely. As journalists who work for publishers, we're already at a crossroads. Do we sign up to such partnerships, or risk being plagiarised anyway? It's a tough position to be in.

To be clear, this isn't sympathy for people or companies plagiarising work. But at the end of the day, you're going to get railroaded either way, if you don't find a way out.

Some of the bigger questions revolve around how organisations want to move forward. Do they want to make these investments in-house or just use third parties, like Perplexity? Do media houses plan to invest in product development? Is editorial and product willing to work together?

Keep in mind that none of these answers are easy, or cheap for that matter. As far as we know, there's no one right answer unfortunately. If you think you've got an answer or a solution, let me know. I'd love to pick your brain about it.

That's a wrap! If you liked what you read, share it with folks you know. If you didn't like something, drop us a mail with feedback. Until next time!

Beyond Tomorrow is a weekly newsletter published every Saturday to give you a roundup of everything AI in the last week.

Essential Business Intelligence, Continuous LIVE TV, Sharp Market Insights, Practical Personal Finance Advice and Latest Stories — On NDTV Profit.