AI ASSISTANTS

ChatGPT vs. Claude: how people really use AI

It would be simplistic to call it the “Battle of the Bots.” But whether it was a coincidence or a calculated attempt at one-upmanship on behalf of either OpenAI or Anthropic, new usage reports on their AI assistants ChatGPT and Claude arrived within hours of one another on Monday.  

As you would expect of two pieces of work published individually, compiled using different datasets and methodologies, “How People Use ChatGPT” and the “Anthropic Economic Index” produced some contrasting results.

However, it was possible to draw up some overarching conclusions, specifically that ChatGPT – used by 700 million people weekly by the end of July – is being engaged more for personal tasks, including writing and research, while Claude is being leaned on for work-related jobs.

Among the key takeaways from the OpenAI data is that the trend for non-work-related usage is increasing, with the number of messages in this sphere growing from 53% of all usage in June 2024 to 70% by June 2025. Although the number of work-related messages rose over the same timeframe, they dropped in percentage, from 47% to 27%. However, OpenAI does point out that the “primary sample” for its dataset were messages sent on consumer plans (Free, Plus and Pro). 

When ChatGPT is used for work-related purposes, “information seeking and decision support” are more common use cases than requests to perform tasks directly. “Educated users and highly paid professionals” are most likely to use it for work.

Contrastingly, Anthropic’s Claude has seen usage by businesses increase massively, particularly in education (which now constitutes 12.4% of all use) and scientific research (7.2%). However, coding continues to lead the way here (36%), while there’s a growing willingness to embrace full autonomy, reflected by a rise in the delegation of entire tasks (up from 27% to 39%).

While there are vast amounts of data from both to dissect and digest, one thing seems abundantly clear: a “Battle of the Bots,” if there ever was one, is becoming increasingly unnecessary as each evolves to fill a different role for users.

Anthropic's and OpenAI's reports highlight a broader trend in AI adoption, the consumer versus enterprise divide. ChatGPT is being used more for personal and creative tasks like writing and research, while Claude is seeing more use at work and in education, coding and scientific research. They have different uses and different fits, but that shows there's room for both, with each AI assistant doing what it does best for the right audience. 

TOGETHER WITH PACASO

Former Zillow exec opens door to $1.3T market

Austin Allison sold his first company for $120M and served as an exec for Zillow. However, both companies reached massive valuations before regular people had the chance to invest, so he built Pacaso differently.

Pacaso has made $110M+ in gross profits to date by disrupting the $1.3T vacation home market. No wonder the same VCs that backed Uber, Venmo, and eBay already invested in Pacaso. 

Unlike his previous companies, you can invest in Pacaso as a private company. Over 10,000 regular people have already invested, but time’s running out. 

Invest before Pacaso’s investment opportunity ends on 9/18.*

AI ON TRIAL

Rolling Stone publisher sues Google over AI use

Penske Media Corp., the publisher of Rolling Stone Magazine and The Hollywood Reporter, has filed an antitrust lawsuit against Google, claiming it is using its content for AI summaries without permission.

The lawsuit against Google and its parent company, Alphabet, was filed in a federal court in the District of Columbia, marking the first by a major U.S. publisher against Google's AI Overviews feature. 

The lawsuit claims Google is abusing its dominance in online search to strong-arm publishers like Penske to supply content Google republishes without permission in its AI-generated answers – those summaries that appear prominently at the top of a web page when you're searching for something.

The suit goes on to say that those AI-generated answers "unfairly compete for the attention of users on the Internet in violation of the antitrust laws of the United States."

"This conduct threatens to perpetuate Google's general search monopoly into the era of generative search and to expand it into online publishing, restricting competition in those markets and reducing the production of original content for consumers," the lawsuit states. 

Penske claims Google's actions are threatening to publishers like PMC, who depend on referrals from Google's search engine for a large portion of the revenue that it devotes to producing original online content through its more than 25 print and digital properties. 

"We are a CONTENT company; content is our lifeblood. We must remain dedicated to delivering the most unique and compelling content to our audiences every day, never sacrificing a strong point of view for a quick win," Penske said in the filing. 

Jay Penske, PMC's chairman, founder and CEO, told Axios that PMC's duty as a global publisher is to protect its "best-in-class journalists and award-winning journalism as a source of truth." 

This isn't the first time Google has faced a publisher in court for illegally using content without permission in its AI summaries. 

In February, online learning platform Chegg filed a lawsuit against Google and Alphabet, claiming it used Chegg's content without permission. Similar to the case involving Penske, Chegg alleges that Google publishes its content in AI-generated answers, which unfairly competes for user attention on the Internet and violates U.S. antitrust laws.

The news comes just a week after Anthropic agreed to pay a record-breaking $1.5 billion to settle a class-action lawsuit from authors over copyright infringement.

TOGETHER WITH BRIGHTWAVE

From data overload to clarity: how Brightwave redefines research

Brightwave's state-of-the-art multi-agent research system is the most powerful synthesis engine on the planet.

10k+ documents, long-running asynchronous background agents, fine-grained context control and a flexible, intuitive UI that thinks like you do.

API access available. Try Brightwave today.

GOVERNMENT

AI bot takes joins Albania’s Council of Ministers

The idea of a government minister responsible for artificial intelligence has gained momentum over the last year or so, with the United Kingdom and France among the European states to make appointments. However, the small Balkan nation of Albania, with a population of 2.7 million, has taken things a step further by appointing the world’s first-ever minister developed by AI.

The virtual entity is known as Diella – the female form of the word for sunshine in Albanian – and although her unveiling could be construed as largely symbolic, there is serious intent behind the move.

As Prime Minister Edi Rama explained via a post on Facebook, the bot will have a key role in taking procurement decisions away from the ministries to ensure that “public tenders are completely free of corruption.”

While the announcement predictably caused unprecedented interest among global media outlets, the appointment of Diella, illustrated by an avatar of a young woman in traditional clothing, was less of a surprise to seasoned Albania watchers. Indeed in August, Rama hinted at the direction of travel, saying: “One day we might have a ministry run entirely by AI. That way, there would be no nepotism or conflicts of interest.”

Diella isn’t entirely new to Albanians, either, having served as a virtual assistant on the e-Albania platform, which helps citizens with public services and issues digital documents, since the start of the year.

Rama’s hope is that by using the latest AI models, Diella can deliver complete transparency in showing how criteria have been met when proposals are submitted online. And there is significant incentive for Albania’s pioneering AI minister to successfully execute this role, with the European Union making clear that progress in combating “corruption and organized crime” will play a key role in Albania’s bid to join the EU.

Over to you, Diella.

LINKS

• S&P 500 marks a new all-time high, closing above 6,600. 

• California lawmakers pass AI safety bill, governor to decide

• Amazon’s fall hardware event is coming Sept. 30.

• SpaceX’s new contracts bring funding close to NASA levels

• Qualcomm announces its latest flagship processor

• OpenAI boosts Codex with GPT-5 upgrade

• Meta accidentally leaks new smart glasses

• Dell Pro Max GB10: AI developer PC with NVIDIA stack, Blackwell chip, and 1 petaflop of FP4 compute*

• Guidde: AI-powered tool to auto-generate step-by-step video guides, docs, and voiceovers with a free browser extension*

• SnowSEO: Streamline optimization, boost rankings and drive organic growth 

• Cal.com: Lifelike agents help book meetings, send reminders and more

• Guidde: Magically create video hot-to guides, training materials with AI

• Scaloom: AI Reddit marketing tool that drives customers on autopilot

• Pinterest: Data Scientist

• Nuro: Senior Software Engineer, Perception, Machine Learning/Computer Vision

• Zensar Technologies: Artificial Intelligence Engineer

• PayPal: Data Scientist

GAMES

A QUICK POLL BEFORE YOU GO