Among the many challenges facing today’s global workforce is concern that there are too few people knowledgeable enough about the opportunities and risks associated with generative AI — that is, people who know how to use emerging tools effectively, pragmatically and ethically to do their jobs, whatever those jobs may be.
So, that’s why there’s been a focus on upskilling today’s workers and putting in place programs for the up-and-coming workforce of students in colleges today. And that’s why the University of Pennsylvania’s School of Engineering and Applied Science announced in February that it was the first Ivy League school to offer an undergraduate major in AI. It followed up last week with news that it’s now the first Ivy League school to offer a graduate degree in AI.
But the Ivies aren’t the only ones working on the problem. Google and IBM, part of a consortium of nine companies, including Cisco, Intel and Microsoft, joined forces last month and announced a new effort with the goal of evaluating “how AI is changing the jobs and skills workers need to be successful.”
What are they doing specifically? Google, through its AI Opportunity Fund, is giving out $75 million in grants to organizations to teach Americans how to use AI, the company said last week. It also launched the Google AI Essentials online course to help teach foundational AI skills and best practices. The $49, 10-hour course, available on Coursera, will teach people how to use AI in day-to-day work. Citigroup will also use the course to upskill its employees, CNET’s Imad Khan reported.
Meanwhile, IBM said it’s already helped train over 11.5 million people through its SkillsBuild initiative and will keep adding new free AI classes. Its latest course, Introduction to Generative AI, is a 1.5 hour class for people who aren’t tech experts but want to get up to speed on the topic. Other classes cover the art of the prompt, and understanding natural language processing.
For people who’ve been let go from their jobs in the past year, online learning platform Udacity said April 30 that it’s offering laid-off workers a free, all access pass to Udacity’s entire catalog of classes (AI and beyond) for 30 days. Go here to register.
Here are the other doings in AI worth your attention.
Ukraine’s new spokesperson is an AI named Victoria Shi
To save time and resources, the Ukrainian government will be using an AI-generated digital person to give official statements and provide updates on behalf of the foreign ministry, according to Agence France-Presse.
The digital spokeswoman is called Victoria Shi, with her name based on the word “victory” and the Ukrainian term for artificial intelligence: “shtuchniy intelekt,” the news outlet reported. “Shi’s appearance and voice are modelled on a real person: Rosalie Nombre, a singer and former contestant on Ukraine’s version of The Bachelor reality show,” AFP said.
“We are taking a step into the future,” Shi says in a YouTube video posted by France Inter, a public broadcast service. “First and foremost, I will inform the public, providing timely and verified information from Ukraine’s consular service.”
I encourage you to watch Shi — kind of eerie.
Ukrainian Foreign Minister Dmytro Kuleba said the new spokeswoman was a “technological leap that no diplomatic service in the world has yet made,” according to AFP.
Shi was created by a group called The Game Changes, which has also made VR content about Ukraine’s war with Russia, the news service said. To avoid people confusing the real fake spokesperson’s statements with bogus copies, the government said, Shi’s statements will include a QR code that links them to text versions on the ministry’s website.
One important note: Ukraine said all of Shi’s statements will be written and verified by humans and not generated by AI.
Apple buys AI company focused on portable devices
At its developers’ conference in June, Apple is expected to provide more details on how it’s using AI in its products and services. And now it’s bought another AI company, according to MacRumors, which cited a European Commission filing reported on by the French publication Challenges.
The startup, called Datakalab and based in Paris, reportedly has technology that furthers Apple’s plans to “deliver on-device AI tools.” MacRumors noted.
“The company was established in 2016 by Xavier and Lucas Fischer and made significant strides in AI technology focusing on low-power, high-efficiency deep learning algorithms that function without relying on cloud-based systems,” MacRumors said. “The startup’s expertise in compressing neural networks to work effectively on portable devices like smartphones and tablets is likely a key factor in Apple’s interest.”
Apple periodically buys companies without disclosing the purchase. That was the case with DarwinAI, a Canadian AI startup that uses vision-based technology to assess components during the manufacturing process, Bloomberg said when it announced news of the acquisition in March. Terms of the deal weren’t announced.
As for Apple’s AI ambitions, we should learn more during the company’s Worldwide Developers Conference, which is set for June 10.
“Multiple reports suggest Apple is in talks with Google to bring its Gemini AI model to iPhones, which could help the devices keep up with generative AI on other mobile devices,” CNET’s David Lumb reported. “Whether Apple has plans to use other generative AI solutions (or its own), the rumors indicate more attention on AI than we’ve seen from the company before.”
Apple CEO Tim Cook also used the company’s May 2 earnings call to address his thoughts on AI, saying, “We continue to feel very bullish about our opportunity in generative AI.”
“We are making significant investments, and we’re looking forward to sharing some very exciting things with our customers soon,” Cook said on the call. He said Apple has advantages “that will differentiate us in this new era” and offered as examples the company’s “seamless hardware, software and services integration,” its homegrown computing chips, its neural engines and its focus on privacy.
Meta will reap big rewards from AI, just not any money for now
During Meta’s earnings call late last month, CEO Mark Zuckerberg said out loud what many investors in AI technology already know: It’s going to take a while for AI investments to make any money (an observation that sent the company’s shares tumbling).
“Historically, investing to build these new scaled experiences in our apps has been a very good long term investment for us and for investors who have stuck with us. And the initial signs are quite positive here too,” Zuckerberg said in the earnings call with investors. “But building the leading AI will also be a larger undertaking than the other experiences we’ve added to our apps, and this is likely going to take several years.”
Meta, which today gets most of its profits from ad sales, shared the complete transcript of Zuckerberg’s remarks on the call here.
The CEO also explained his plan to reap profits from the AI spend. “There are several ways to build a massive business here, including scaling business messaging, introducing ads or paid content into AI interactions, and enabling people to pay to use bigger AI models and access more compute,” Zuckerberg said. “And on top of those, AI is already helping us improve app engagement, which naturally leads to seeing more ads, and improving ads directly to deliver more value.”
TL;DR: “If the technology and products evolve in the way that we hope, each of those will unlock massive amounts of value for people and business for us over time,” Zuckerberg added, noting that “tens of millions of people” have tried its new Meta AI assistant, which was released in mid-April on its popular Facebook, Instagram, WhatsApp and Messenger platforms.
Of course, “if” is the operative word, given that Meta is in competition with OpenAI, Microsoft, Google, IBM and others to win over consumers with its AI offerings.
Scientists spot audio deepfakes by looking for ‘signs of life’
Klick Applied Sciences, which made news after using 6-to-10-second smartphone voice recordings to create an AI model that helps identify people at risk for Type 2 diabetes, has now turned its attention to helping spot audio deepfakes.
Basically, Klick Labs said it’s using AI to look for “what makes us human,” as part of its effort to separate real audio from deepfake audio of everyone from President Joe Biden to singer Taylor Swift.
“Klick researchers created an audio deepfake detection method that taps into signs of life, such as breathing patterns and micropauses in speech,” the researchers wrote in a paper called “Investigation of Deepfake Voice Detection using Speech Pause Patterns: Algorithm Development and Validation.” The paper was published in March in the open-access journal JMIR Biomedical Engineering.
The Klick Labs team said it studied 49 people from diverse backgrounds and with diverse accents. “Deepfake models were then trained on voice samples provided by the participants, and deepfake audio samples were generated for each person. After analyzing speech pause metrics, the scientists discovered their models could distinguish between the real and fakes with approximately 80 percent accuracy.”
Added Yan Fossat, senior vice president of Klick Labs and principal investigator of the study: “Our findings highlight the potential to use vocal biomarkers as a novel approach to flagging deepfakes because they lack the telltale signs of life inherent in authentic content. These signs are usually undetectable to the human ear, but are now discernible thanks to machine learning and vocal biomarkers.”
Well, I’m certainly listening. I hope the US Federal Communications Commission is too — in February, it banned the use of deepfake voices in robocalls, after fraudsters created a phony message from Biden and delivered it to voters in New Hampshire to disrupt the primary-vote process.
More media companies sue OpenAI and Microsoft over copyright
The New York Times generated headlines about itself in December when it announced it was suing OpenAI and Microsoft for copyright infringement, alleging that the companies used Times stories without permission to help train their generative AI engine. (In March, OpenAI and Microsoft filed motions to have parts of the NYT suit dismissed — I have links to all the filings here.)
In February, The Intercept, Raw Story and AlterNet also sued OpenAi and Microsoft, which has invested more than $13 billion in ChatGPT maker OpenAI and uses its technology to power AI offerings including Microsoft Bing.
Now we can add eight other news outlets to the list of copyright owners going after OpenAI and Microsoft, including the Chicago Tribune and the New York Daily News, according to Reuters. The papers are owned by investment firm Alden Global Capital’s MediaNews Group, which has accused the tech companies of using its reporters’ work to train their AI systems.
“The lawsuit said Microsoft and OpenAI’s systems reproduce the newspapers’ copyrighted content ‘verbatim’ when prompted,” Reuters reported. “It said ChatGPT also ‘hallucinates’ articles attributed to the newspapers that harm their reputations, including a fake Denver Post article touting smoking as an asthma cure and a bogus Chicago Tribune recommendation for an infant lounger that was recalled after being linked to child deaths.”
OpenAI told Reuters that it takes “great care in our products and design process to support news organizations.”
Some publishers have reached licensing deals with OpenAI, Nieman Labs noted, including the Associated Press, Le Monde and Axel Springer, publisher of Insider.
Last week, The Financial Times said it signed a licensing deal with OpenAI. “The new deal will allow ChatGPT to pull information in real time from the FT’s published stories when answering user prompts. Details taken from those stories will appear in ChatGPT as a summary or quote, with a link to the specific article being cited,” Nieman Labs reported.
Couple saves $10,000 on wedding plans with AI
ChatGPT, the most popular gen AI chatbot in terms of user engagement, can now add wedding planning to the list of things it can handle for you.
A couple in New York used OpenAI’s chatbot to help plan their August wedding and said ChatGPT saved them $5,000 to $10,000, according to NBC News. Maria Cortese and her fiance said they got suggestions on everything from flowers to an affordable photographer to save-the-date invites (with wording generated by the AI), eliminating the need for a wedding planner. Cortese said the chatbot also helped them find deals on goods and services.
NBC then asked ChatGPT for the best destinations for weddings (Hawaii, Napa Valley and New York are popular choices) and to share info on the most hated songs for weddings (ChatGPT said that YMCA by the Village People is disliked as being “cheesy and repetitive”).
According to Zola, a wedding planning site, the average price for a wedding in 2024 will be about $30,000. In a survey of 7,000 couples getting married this year, Zola found that tech in one form or another is used in wedding planning in a big way. It said it found “86% of couples making choices based on social media” and added that 7% of couples admitted to already using at least one AI tool, with another 11% planning to do so.
And yes, that surely includes using AI to help write their wedding vows. No judgment here.
Editors’ note: CNET used an AI engine to help create several dozen stories, which are labeled accordingly. The note you’re reading is attached to articles that deal substantively with the topic of AI but are created entirely by our expert editors and writers. For more, see our AI policy.
