OpenAI Is Introducing New Models But How Much Is Too Much?

OpenAI CEO Sam Altman just debuted the company's latest AI agent dubbed Deep Research.

The full 03 reasoning AI model is sure to give rise to an autonomous assistant that is designed to scour online pages and other digital scholarly sources for data regarding a certain issue. The agent can comply with new reports while users go about their business in different tabs.

You can even leave your PC behind as a whole, allowing the agent to produce several minutes or a few hours with alerts. Unlike Google’s Deep Research, the value of OpenAI 03 Deep Research was witnessed by many in the outside AI world.

It was first shared as a product that only those with the ChatGPT Pro subscribers could benefit from. However, this time around, it’s going to be up for grabs to free and Pro users.

Obviously, those getting it for free have limited uses as compared to those paying for the subscription. While 10 does seem like a reasonable amount, it all depends on how well the agent works. The goal is to encourage more people to upgrade to higher-cost plans to avail more benefits.

If you happen to be a free ChatGPT user, two uses per month are limited to queries or things you’re in search of with Deep Research. However, the company should hope the latest 03 Deep Research launch is worth that price tag.

Next, Sam Altman is clearing the air about how AI models are not going to shy away from sensitive categories. They will refrain from making assumptions that seem out of their view. In a new and updated version of the Model Spec, the company shared how the models will not attempt to steer users in the pursuit of a better agenda, either through direct or indirect means.

The company strongly feels in matters like having intellectual freedom that entails freedom to hear and discuss concepts. This is why Sam Altman is promoting greater transparency and avoiding certain viewpoints that could be shunned forever. It might also have to do with the growing amounts of political pressure in everyday life.

Many of the current President’s closest allies such as Elon Musk and David Sacks went out to accuse AI agents of censoring data linked to conservative views. Some accused the company of being too woke and unreliable about political subjects.

OpenAI says it has major plans to make its models simpler than before as it tries to ship non-reasoning models like GPT-4.5. This is usually a very rare admission from a tech organization. After all, most of these product releases are not differentiated enough for various clients.

This is why Altman just took to the X app to mention new updates on the product and acknowledge so many of the organization’s recent releases. They wish to do some better jobs of sharing and simplifying. They would like AI to work and not complicate models which so many previous designs have been accused of.

The number of models that users can select totally depends on the subscription tier. Altman explained how the firm would like to combine its technology including the o-series of models into the latest GPT-5 when it comes down to APIs and ChatGPT. For the past one year, OpenAI shared its GPT-40, o1, o3, and o3-mini models. It also rolled out the latest ChatGPT subscription tiers such as the $200 per month ChatGPT Pro subscription.

Above all, the organization shared its Operator agent called Deep Research where it tasks ChatGPT and a few other new features that aren’t available via specific designs or paid levels.

Image: DIW-Aigen

Read next: New Study Shows ChatGPT Doesn’t Use as Much Energy as Most Assume
by Dr. Hura Anwar via Digital Information World

New Study Shows ChatGPT Doesn’t Use as Much Energy as Most People Assume

According to a new study from Epoch AI, ChatGPT doesn't consume as much power as many people think. The energy ChatGPT consumes depends on how it is being used and what model of ChatGPT is answering the queries. The estimates from most sources say that answering one question with ChatGPT requires 3 watt-hours of power which is 10 times more than what it takes for Google search to answer one question. But Epoch AI says that ChatGPT using 3 watt-hours of power for one query is too high and OpenAI’s GPT-4 model only takes 0.3 watt-hours to complete one query which is even less than the power used by household appliances.

The environmental challenges with AI’s energy consumption have become a heated topic because AI companies are continuously expanding their infrastructures. Recently, about 100 organizations have signed an open letter which is urging the industry to ensure that new AI data centers which are being built do not harm the environment. So the industries are asked to rely on non-renewable energy sources.

The researcher on the study, Joshua You, said that there is a lot of discourse about how much energy is going to be used in the future but there hasn't been conversations which accurately portray AIs energy consumption right now. He said that AI using 3 watt-hours per query seems too high and it was based on some of his older research and rough calculations. OpenAI hasn’t shared how much its models use energy per query but You says that 0.3 watt-hour seems like an accurate figure. The analysis hasn't calculated the extra energy used by ChatGPT like for processing longer inputs or image generation.

AI is getting more advanced and it is going to require more energy for handling complex tasks in the future. By 2030, it is predicted that training cutting-edge AI models will require 8 GW of energy which is equal to eight nuclear reactors. Joshua You says that people who are concerned about AI’s energy footprint should stop using AI models like ChatGPT or opt for smaller models like GPT-4o-mini.

Read next: AI Chatbots Drive Higher Engagement and Transactions Than Search Traffic
by Arooj Ahmed via Digital Information World

AI Chatbots Drive Higher Engagement and Transactions Than Search Traffic

According to a new study of over 7 million sessions, conducted by Kevin Indig, AI chatbot traffic from platforms like CoPilot, ChatGPT and Perplexity converts better than traditional search traffic. The data which was analyzed was mostly from e-commerce purchases and software sign-ups. Even though traffic from AI chatbots is still less than 1% of total organic traffic, it is growing quickly and will play an important role in future digital marketing.

The study found that 80% of the transactional traffic to the top 10 landing pages goes to sites like eBay, Amazon and Walmart, which are dominating Google Search as well as AI chatbots. Microsoft also shows up often on AI chatbots because of people searching for software like Teams or CoPilot. The sites that are dominating search results are also leading in AI but AI chatbots are becoming more conversational, meaning that ranking factors are going to emerge which will allow LLMs to determine their own ways of visibility.

The traffic on AI chatbots is also more engaged than search traffic, measured by session duration and number of page views per session. However, referrals on AI chatbots are longer than Google traffic like CoPilot which has an average session duration of 11.6 minutes. On the other hand, AI chatbots have an average duration of 10.4 minutes as compared to Google which has 8.1 minutes of session duration. AI chatbots are also gaining transactions with Perplexity and CoPilot seeing bigger increases in pageviews than Google. Even though CoPilot has not as many page visits as Perplexity, it is still showing better engagement as it is integrated in all of Microsoft’s devices and software.

The study also looked at how deep the traffic is going into a website as AI chatbots are also sending more traffic to homepages. ChatGPT sends 30% of its traffic to home pages while Google sends only 10%. AI chatbots are also sending traffic to deeper pages than Google which is sending visitors to second as well as third-level pages. This shows that AI chatbots are driving more engaged traffic which is spending more time on sites. This is also leading to more viewing on pages, making them better potential for conversions. This raises the question whether AI chatbots can improve user engagement by directing users to more specific pages instead of home pages. But it is too soon to say anything because search features of AI chatbots are still under development and will take a long time to get trained over most of the web.

Read next:

• Security Researchers Engineer New and Safer Form of QR Codes That Protects Smartphones From Attacks

• AI Faces Skepticism in Moral Decision-Making Despite Advancements, New Study

• 71.5% Use AI for Search, Yet Only 14% Rely on It Daily as Traditional Engines Dominate
by Arooj Ahmed via Digital Information World

71.5% Use AI for Search, Yet Only 14% Rely on It Daily as Traditional Engines Dominate

According to a new survey by Higher Visibility, most users are using AI just as a complementary search engine and not a replacement, with only 14% of the respondents using AI tools for search daily. Overall, 71.5% of the respondents said they use AI tools for search, with 79.8% reported using Microsoft Bing and Google for informational searches. 20.2% of the respondents of the survey said that they have changed their primary search engines in the last year. Even with people using AI tools for search, traditional search engines are still dominating but SEO is becoming difficult with click through rates on Google declining everyday. People are switching to AI tools and social media sites for searches and they are playing a role in people’s search habits.

The survey also found out AI search trends amongst different generations, with Gen-Z being the generation to use AI tools for search the most. 82% of the Gen-Z have been using AI search tools every now and then and they are also using social media for product discovery. Millennials are using AI tools as well as traditional search engines, while 65% of the Gen-X are also using AI tools every now and then but still prefer traditional search engines. Baby Boomers are using AI tools the least (45%) for search engines but they still use the traditional search engines the most.

It was also found that people are using different types of platforms for different purposes like Americans prefer traditional search for general information but use Microsoft Bing and Google for initial product research. AI tools like Claude, ChatGPT and Bard are being used by respondents for product recommendations and comparisons while discovery based shopping is preferred on social media sites like Instagram, Facebook and Nextdoor. Specialized product searches are mostly done on platforms like eBay and Etsy. Local business searches are mostly preferred on Google Maps, Social Media, Business Profiles and Review Sites.

Read next: Gemini, ChatGPT, DeepSeek: The Biggest AI Data Collectors Revealed
by Arooj Ahmed via Digital Information World

New Survey Shows What Most People Think of AI and AI Generated Content

The researchers from Rutgers University conducted a study to find out people’s view of AI and AI-created content in their daily lives. The analysis found that most Americans have mixed feelings about AI so the researchers wanted to understand the opinion of people on AI in their everyday life. It was found that 47% of Americans trust AI to benefit the public as compared to 39% who said the same about social media and 42% who said the same about Congress. 52% of men and 43% of women trust AI and younger adults between the ages of 25 and 44 trust AI the most (55%). Urban residents (53%) trust AI more than people living in rural areas (38%).

The research also tried to understand how much people trust businesses who are using AI and it was found that 50% trust businesses to use AI responsibly. This trust was seen the highest among people with graduate degrees and who are earning more than $100K (65%). 53% of the people living in urban areas trust businesses who are using AI as compared to 42% of people from rural areas. 48% of the respondents also trust news generated by AI while 62% trust mainstream journalists, showing that people trust human journalists over AI generated news.

Young adults said that they are more confident in recognizing AI generated content, with 13% of overall respondents saying that while 30% feel only somewhat confident. 26% also said that they are very familiar with AI as compared to 63% who do not feel very familiar with it. Men, young adults and educated people are the most familiar with AI. When respondents were asked to take an assessment on their knowledge for AI, the average score was 3.3 out of 8 with people with higher education getting the highest scores.

Read next:

• Gemini, ChatGPT, DeepSeek: The Biggest AI Data Collectors Revealed

• AI Boosts Productivity But At The Cost of Eroding The Human Mindset

• 36% of Jobs Use AI for 25% Tasks, 57% Augment Work, 43% Automate, Software Leads at 37.2%
by Arooj Ahmed via Digital Information World

36% of Jobs Use AI for 25% Tasks, 57% Augment Work, 43% Automate, Software Leads at 37.2%

The Anthropic Economic Index which was released this week provides detailed analysis about industries that have been using AI by using the data from Claude. The report also talks about that AI isn't going to replace majority of jobs anytime soon but it is changing how we work in workplaces by providing help in areas like technical writing, software developer and business analysis and 36% of the jobs have been using AI for at least a quarter of their tasks. The report focuses on finding real world usage of AI and its adoption instead of using predictions and surveys by analyzing four million conversations of users with Claude.

The findings of the analysis shows that 57% of the AI usage is augmented which helps AI in tasks like polishing ideas, brainstorming and checking accuracy of works. On the other hand, 43% of AI usage is direct automation in which AI handles tasks with minimal human help. This shows that businesses are using augmentation as well as automation to do their tasks. It was also found that AI isn't replacing the jobs entirely, it is just being used for assistance in specific tasks. 4% of the jobs are using AI for 75% of their tasks which shows that AI can be helpful in certain roles. 36% of the jobs are using AI for 25% of the tasks which tells that people are slowly integrating AI to become a part of their jobs.

The report also shows in which fields AI adoption is strongest, with fields kike media, tech and business being at the top. The analyzed conversations with Claude shows that software development is the leading field with AI interactions (37.2%) and people are using it for tasks like modifying software, debugging codes and troubleshooting networks. 10.3% of the analyzed conversations were about creative and editorial roles where people are using AI for researching, drafting text and generating ideas. AI adoption was found to be lower in jobs with physical labor like transportation, healthcare and agriculture. Only 0.1% of the conversations involved fishing, farming and forestry tasks. This shows that most people are using AI in text based and analytical tasks while it is not being helpful to people in physical work and complex human interactions.

An interesting thing from the report was that AI adoption doesn't follow a certain wage pattern, and it is higher in mid to high salary ranges. The report says that AI usage is higher in the upper quartile of wages mostly in software roles, but is lower in very high wage jobs like physicians and very low wage jobs like restaurant workers. This shows that AI adoption and usage doesn't depend on highest levels of expertise or wages, rather it depends on how much technical and analytical skills are required for a job. The report offers a helpful guide for business leaders who want to know about AI’s impact and how its adoption will bring a change.

Read next:

• QuillBot: The Game-Changer for Enterprise Customer Support Teams

• Gemini, ChatGPT, DeepSeek: The Biggest AI Data Collectors Revealed

• AI Boosts Productivity But At The Cost of Eroding The Human Mindset
by Arooj Ahmed via Digital Information World

AI Boosts Productivity But At The Cost of Eroding The Human Mindset

A new research featuring more than 300 knowledge workers is taking center stage as we speak and it’s talking about the use of AI in our daily lives.

The study published at CHI-25 shares more about how professionals from different fields are using generative AI to carry out routine tasks daily. The results prove that while AI does boost productivity, it can erode important cognitive skills.

There is a lot of confidence linked to Generative AI but that means humans are engaging less in critical thinking. Experts don’t feel that’s healthy because high self-esteem only comes with more you think critically. So relying on this technology might be doing your personal self little good.

The research experts call such a dynamic a major shift from producing materials to critical integration. This is where workers are spending much less time making content through scratch but they need to add in more efforts to better refine and verify the outputs generated through AI.

There are many experts out there today and so many AI tools that become such an important part of your life and the working world. As per the research, workers use AI assistants for all things such as writing codes and data analysis and drafting their own emails. They’re now making presentations through such means and about 97% of all those involved reported utilizing ChatGPT. Other tools such as Copilot by Microsoft and Gemini by Google are very popular.

As per the study, there are three main shifts in terms of how a professional engages with work when using AI. Some data gathering is transformed into data verification. Then problem solving switches to response integration and then task execution entails task stewardship.

One of the study’s participants highlights more details about the different challenges with AI including verification difficulties. It takes a lot of time to verify manually and the same goes for users in different professions. They need to keep on cross-referencing AI outputs against authentic sources.

One major factor why professionals rely on AI is limited time. They prefer high speeds over verification. If the tech is helping them save time, there’s nothing better and they don’t need to worry about the final results.

All researchers warned that such behavior patterns might give rise to cognitive atrophy. In other words, it greatly diminishes your mental capabilities when you start to rely more on AI to get the task done. Such a phenomenon is a reflection of concerns from the past about their personal struggles with something.

The research also noted that workers having greater confidence in themselves and their talents could engage with AI more critically. They would use tools to assist with judgments and not replace them entirely. These people were spending more time on evaluations and refinement of content made using AI, especially for tasks needing domain expertise.

So the take-home message from this study by researchers is companies need to create strategies to ensure employees’ critical thinking skills remain intact, with or without using AI. They recommend curating AI tools that market critical engagement instead of simple output acceptance blindly.

Image: DIW-Aigen

Read next: This Billionaire Battle Just Got Personal, Musk vs. Altman Is the AI Drama of the Year!
by Dr. Hura Anwar via Digital Information World