Journey of AI

An AI for an AI

Industry over taking academia in AI

In 2023, the industry produced a notable of 51 AI models while academia contributed just 15.
In 2014, when universities led the AI research, according to Stanford University's 2024 AI Index report.

Trump dilutes AI regulations

US President Donald Trump returned to white house in january 2025, he swiftly dismantled AI regulatory guardrails, indicating that corporate ambition and not government caution would dictate AI's future.

Trump alliances with big tech

Made alliance between the US government and its tech giants and to keep China out of the race for AGI.

Stargate

Trump announced the largest AI infrastructure project in history named , Stargate, with US $500 billion investment in AI infrastructure and colossal data centres across US.
This was to re industrialise the US.

Stance of big tech

OpenAI criticised the attempts by individual US states to introduce their own AI regulations.
Google adopted a pro innovation approach framed around national security. In its proposal for AI action plan, google urged the US administration to combat restrictive foreign AI barriers, warning that some governments were imposing regulations that could disproportionately affect American companies.

EU suing US tech firms

At the World Economic Forum on january 23, 2025, Trump condemned EU's treatment of US tech firms, relating to the cases against Apple and fined it for $15 or $16 billion and now after Facebook.
["] "Whether you like them or not, they are American companies and they should not be treated that way" - Donald Trump.
The European Commission has fined Google $2.66 billion for breaching anti trust rules in 2017 and imposed $1.32 billion fine on Meta in 2023 for transferring European data to the US in violation of data protection laws.
At the Paris AI Action summit, 2025, US VP JD Vance reinforced the administration's stance, "The Trump administration is troubled by reports that some governments are looking to clamp down on American AI companies and America cannot and will not accept that. "

A strong lobbying efforts by big tech ad member states succeeded in watering down much of the EU's AI act, particularly regulations around generative AI.

Germany, Italy, France sidelining with big tech

EU countries reached a political agreement in october 2023, Germany, Italy, France coordinated efforts to remove most provisions on generative AI, threatening to vote against the entire act is no changes were made.
Shortly after the final agreement, Mistral AI announced a partnership with Microsoft in february 2024.
66% of AI related meetings involving European parliament members were with corporate interests, up from 56% in 2019 - 2022 - a research by corporate Europe observatory, research group.

China's AI development

In China, private companies are driving majority of AI development. China's AI environment is more diffused. Several startups like Zhipu AI, MiniMax, Baichuan AI, DeepSeek are leading large language model developments, backed by tech investors and venture capitalists.
In january 2025, DeepSeek released a model matched OpenAI, but at just $6 million in costs, a fraction of $78million that OpenAI has spend in training its chat bot GPT-4, according to 2024 AI Index report.
2 months later, chinese tech giant Baidu made headlines with 2 low cost AI models that claimed superior language skills and logical reasoning abilities at just 1% cost of OpenAI's costs.

Lavender AI by Israel

In 2024, article on Israel's use of digital tools against Palestine, "Lavender", an artificial intelligence tool that assigns Gaza residents a numerical score relating to suspected likelihood that a person is a member of an armed group. If the scores exceed a threshold, the individual may be targeted.

Autonomous Weapon system (AWS)

Countries like China, Israel, Russia, South Korea, Turkiye, UK and US are investing in building autonomous weapons system (AWS), said in the 2024 article by Carnegie endowment for International peace, a think tank in US.
The US department of defense defines AWS as systems that once activated can select, engage targets without further intervention from a human operator.
Concerns about the AI based weapon going rogue, algorithms built from a biased and incomplete data.
Israeli military designated palestinian human rights organisation as terrorist groups, could increase the possibility that Lavender could target civilians too.
This also evokes the privacy concerns because AI is trained using personal data.

Treatment recommendation based on patients personal details

The study from New York, shows that generative AI models can recommend different treatments for the same medical condition, based on the patient's socio economic or demographic background.

2025 AI Index report, released by Stanford Institute for Human Centered Artificial Intelligence, highlights concerning trend of crimes associated with AI rising 57% between 2023 and 2024.
Deep fake intimate images and chatbots allegedly implicated in a teenager's suicide.

Violating IPR

Claims of AI violating intellectual property rights, in february 2025, some 14 publishers including the The Atlantic, The Guardian Forbes sued Canada based AI developer Cohere Inc. claiming that the company used the publisher's copyrighted works to train its generative models and displayed the news article without attributing to them. They also alleged that the model provided false or hallucinated information and attributed it to them.

Can generate false text

AI systems can generate false text and there can be a case of AI misinterpretation.
AI is found to display traits of deception and manipulation.

Governance of AGI is needed

2024 report by the Stanford Cyber Policy centre says governance of generative AI requires policy makers to walk the line between the enabling AI benefits while warding off the risks.

Japan guidelines

Japan has non binding guidelines applicable to AI development and use and promotes voluntary efforts by AI stakeholders for safety, fairness and transparency.

UK 10 year plan - 2021 - 2030

UK in 2021, introduced a 10 year plan to become a global AI superpower. In january 2025, it launched the AI opportunities action plan.

EU's AI act, 2024

EU, an AI act entered into force in August 2024, the first set of rules implemented in february 2025, which prohibits 8 practices.
In May 2024, Council of Europe, a human rights organisation, adopted the first international AI treaty. The Framework Convention on Artificial Intelligence and Human Rights and Democracy and Rule of law. This has included 13 signatories so far including the EU, US, UK and Israel. The treaty will enter into force after at least 5 signatories including the 3 council members ratify it.

Artificially yours

Jobs getting redundant

Many graphic designing jobs have already become redundant in the past few years.
AI - technology that provides machines the ability to engage in cognitive activities such as reasoning, learning, problem solving in a way similar to human brain but at a faster rate and without fatigue.

Since 1960s

Generative AIs have been in existence since 1960s , but not like ChatGPT, or Microsoft's Copilot, Meta's Llama, Google's Gemini.

Future of jobs report, 2025 - WEF

The future of jobs report 2025, by the World Economic Forum (WEF) states that 2/5th of worker's existing skill sets will be transformed or become redundant over the 2025 - 2030 period, resulting in widening inequality.
69% of the surveyed employers plan hiring new people with skills to design AI tools and enhancements appropriate for organisation specific skills.

UNCTAD report

Report by UN Trade and Development (UNCTAD), around a third of developing nations have AI strategies.
UN Trade and Development (UNCTAD)'s Technology and Innovation report, 2025 also states the same, following the release of ChatGPT, image creation jobs (graphic designing and 3D modelling) saw a 17% decline.

2023 research paper

A 2023 research paper by Imperial college London Business school, UK and Harvard Business School, US and German Institute for economic research, Berlin, quotes that this reshaping of workforce due to AI could happen in 2 ways.

Humans increasingly collaborate with machines improving speed and quality of work. Automating routine and free up workers for tasks that require more from human attention
Technology will create new jobs, roles in AI research and development, AI trainers, AI explainers, AI sustainers (who monitors and refine AI uses)

2025 global survey

2025 global survey on the state of AI by US based consultancy McKinsey and Company, 78% of respondents say their organisation use AI in at least one business function in 2024 up from 55% in 2023.

Ghibli garnered 1 million users

OpenAI's 4O image generation tool launched in March, 2025 garnered 1 million users subscribed within 1 hour, where AI convert into artwork of Ghibli studio, a Japanese animation studio.

AGI

The next focus is on AI agents and Artificial General Intelligence (AGI)
AGI is still in conceptual stage, which can develop an understanding and adapt to a developing situation.
An AGI powered self driving vehicle while encountering traffic congestion, could decide on an alternative route instead of following a pre programmed route. It would communicate with other connected vehicles and change its path considering distance and travel time.
2025 survey by the Association for Advancement of AI, a washington based non profit scientific society, states 76% of leading AI scientists believe that the current machine learning based approach to yield AGI is unlikely or very likely to succeed.
Elon Musk predicts that AGI will become reality in the next 2 years.

Recommendation to Manhattan like project

US China commission, an independent agency of the US government, recommended that the Congress establish and fund a Manhattan project like programme dedicated to racing to and acquiring AGI.

AI agents

Agents carry out multiple tasks with minimal human interventions, acting as a virtual co worker.
Companies like Butterfly Effect, Google, Microsoft, Amazon, Apple, Meta, OpenAI have already developed AI agents.
In 2025, the world may see the first AI agents join the work force - Sam Altman.

There is this tendency in the AI community to equate success with first 90% of progress. But in most real world scenarios, the last 10% is way harder and usually required more tailored solutions - Sayash Kapoor, Princeton University.

The heart of battle are semiconductors.

US industry leads the global semiconductors market, close to 50% of the market share, designs the world's most advanced semiconductor technology.
Other major players in terms of market share are Korea (9%), Taiwan (8%) and China (7%)
In october 2022, the US first imposed restrictions on the sale of advanced semiconductors and equipment used to make semiconductors to China to protect the US national security.
In 2024, it imposed new restrictions, prohibiting export of 24 types of semiconductor - manufacturing equipment, three software tools and other semiconductor technologies to 140 chinese chip making companies.
In 2025, US added 80 entities from China, UAE, South Africa, Iran, Taiwan, and other nations to the list.

The US has divided the world in 3 different groups.

Closest US allies - 18 countries - unlimited access to the semiconductor chips.
Other countries - limited access
Dictatorial or Authoritarian regimes - China, Russia, Iran, North Korea and others - not have any access

China

China, in december 2024, enacted an export ban on critical minerals like antimony, gallium, germanium to the US, needed to make semiconductors.
China maintain 60% of the world's rare earth mining production and 90% of processing and refining, as per International Energy Agency.
China's AI technology has advanced quite well. January 2025, Chinese firm DeepSeek released R1, a LLM developed at a fraction of cost of OpenAI o1, but similar to its performance.

India needs a update

IndiaAI Dataset platform, 2025

In march 2025 the government launched IndiaAI Dataset platform, reducing the barrier to innovation by providing access to data.

AI Compute Portal

It also launched AI Compute Portal to create infrastructure.

AI Mission, 2024

Union cabinet's approval of India's 10,372 crores AI Mission in 2024, centred around 7 verticals.

Computing Infrastructure Capacity
Skilling
Innovation
Datasets
Startup financing
Application development
safe and trusted AI

Draft Digital Personal Data Protection Rules, 2025

The approach raises concerns over data protection under the Draft Digital Personal Data Protection Rules released in january 2025 to operationalise the Digital Personal Data Protection act, 2023.
There are questions of whether leaked personal data available on the internet will be considered publicly available data that can be scraped and used to train a model.

Concern in tax administration

The other major concern is that the government is already embedding AI into their systems without proper guardrails, like the use of AI in tax administration was incorporated under the Taxation (Amendment) Act, 2020 and the amendment to the criminal Procedure (identification) act, 2022.

Issue of increased digitisation

The issue of people being denied access to social scheme due to increased digitisation.

Poshan tracker

Poshan tracker, a tool to ensure real time monitoring of supplementary nutrition provisioning for improving the nutritional status of beneficiaries, has incorporated facial recognition system to streamline the Take Home Ration distribution. "how do you expect pregnant women to walk all the way to anganwadi centres to give facial recognition for take home ration?"

Committee to draft a report

In november 2024, court directed Ministry of Information and Broadcasting & Ministry of Electronics and Information Technology to formulate a committee to draft a report within 3 months, addressing legislative, technical and awareness aspects of deep fake management.
In march 2025, the ministry submitted its report, announcing the creation of the member committee. Asked for another 3 months to submit the report.

Outline for Digital India Act

The government is contemplating a new act to address the growing challenges posed by AI. In march 2023, MEiTY released an outline for the Digital India act to replace the existing Information and Technology act 2000.
The outline suggests that the act will look at the monopoly of big tech companies, encourage growth of startups, address user harms, such as revenge porn, catfishing, doxxing, cyberstalking and phishing and try to reduce hate speech and misinformation on the internet, as well as in the AI.
MEiTY also suggested that the Digital India act could potentially eliminate the immunity granted to online intermediaries, from a legal liability for third part content posted on their platforms under the IT act, 2000.
The government has not publicly released the first draft of the act yet.

Regional spread of AI

Key terms

Artificial Intelligence (AI)

products and services that are adaptable and autonomous. The adaptability means after being trained, often developing the ability to perform new ways of finding patterns and connections in data that are not directly envisioned by human programmers. The autonomy refers that the systems can make decisions without intent or on going control of human.

AI agent

Are software systems that use AI to pursue goals and complete tasks on behalf of users. They show reasoning, planning and memory and have a level of autonomy to make decisions, learn and adopt. ex: China's Manus AI, OpenAI's Operator

AI assistant

a software application that uses AI to perform tasks. It uses natural language processing and machine learning to understand requests and provide response. Unlike AI agents, which are proactive and work autonomously to achieve a specific goal by any means at their disposal, AI assistants perform tasks at request. ex: Amazon's Alexa, Microsoft's Cortana, Apple's Siri

Artificial General Intelligence (AGI)

Sometimes known as general AI, strong AI or broad AI, the terms refer to a theoretical form of AI that can achieve human level or higher performance in most cognitive tasks.

Artificial Neural Network

A computer structure inspired by the biological brain, consisting of a large set of interconnected computational units "neurons" ex: Google's search algorithm

BIAS

Bias in AI models refers to output errors caused by skewed training data.

Deep learning

A subset of machine learning that uses artificial neural networks to recognise patterns in data and provide a suitable output, for example, a prediction. Deep learning is suitable for complex learning tasks such as voice and image recognition, object detection and autonomous driving.

Foundation Models

A machine learning model trained on a vast amount of data

January 2024

40% of global jobs face risks from AI

According to IMF's latest report, almost 40% of all jobs around the world are exposed to AI, with advanced economies at greater risk.

Impact different countries differently

Impact of AI is likely to differ significantly across countries at different levels of development or depending on economic structures.
Emerging markets like India, Brazil, Colombia, South Africa and low income economies may face fewer disruptions compared to advanced economies such as UK and USA.

60% of jobs to be impacted

In contrast to the nearly 60 percent of jobs that may be impacted by AI in advanced economies, the IMF pegged this figure at 40 percent for emerging markets and 26 percent for low-income countries.

High exposure occupations in US and UK

Almost 70 percent and 60 percent of UK and US employment, respectively, is in high-exposure occupations, approximately equally distributed between those that are high- and low-complementarity positions.

High exposure occupations in India, Brazil

High exposure employment in emerging market economies ranges from 41 percent in Brazil to 26 percent in India

Opportunity for emerging markets

AI may also represent an opportunity for emerging markets and developing economies to address skill shortages, especially in the health and education sectors, potentially increasing inclusion and productivity.

Need to understand

The report emphasised that understanding which groups are most vulnerable is “essential to design policies that can mitigate those effects”. Print

July 2025

Trump hails $90 billion in AI Infrastructure Investments at Pennsylvania summit

The event was organized by Senator David McCormick, Republican of Pennsylvania, who brought together Trump administration officials and executives from technology and fossil fuel companies, including Amazon Web Services, Anthropic, Google, ExxonMobil and Westinghouse.

Blackstone's investment

The private equity firm Blackstone announced that it would invest $25 billion in new data centers and energy infrastructure, including natural gas power plants.

Google's investment

Google said it would invest another $25 billion in data centers and announced a separate $3 billion plan to upgrade two of Pennsylvania’s existing hydroelectric dams to produce more electricity.

CoreWeave's investment

CoreWeave, an A.I. cloud company, said it would invest $6 billion in a large data center near Lancaster, Pa.

Want easier rules

Officials have also said they want to make it easier to approve new natural gas and nuclear power plants to supply the enormous quantities of electricity needed to supply data centers

National energy emergency

On his first day in office, Mr. Trump declared a “national energy emergency,” saying the country did not have enough power to meet its growing needs for A.I. and ordering agencies to roll back environmental rules.

Lifting restrictions on selling chips to China

On Monday, the chipmaker Nvidia also said that the administration had lifted restrictions on selling certain types of A.I. chips to China. NYT

February 2026

Savita Bhabhi - AI rewriting desi desire

Savitha Bhabhi, adult comic book arrived in India in the late 2000s, the country's first porn star even without being real.

Her successors, daisy virtual girlfriends and AI wives, neighbor aunties and office colleagues, shy brides and dominant older women, all available and customizable. Now the scene spills into a chat window where the user can step inside. newly typed bots continue to circulate.

Nandini A.I. Bhabhi, a vibrant and relatable daily life cooking home meals and strolling through markets Maya Leela on Instagram, with 95,000 odd followers, feeds in blouseless saris and intimate angles that tease without tipping into the explicit. He is a male architect in Hyderabad who began tinkering with AI tools two years ago and insists this was never meant to be about sex.

Erotic content became a deliberate strategy and the fastest way to test whether a persona feels alive to people rather than abstract or gimmicky. A significant section of the audience doesn't realize she is AI-generated, says the Maya's creator. That uncertainty: is she real? Is she not?

Hypersexualized AI women appear in everyday Indian scenes:

cooking
sweeping
teaching
standing in crowded canteens
or in buses where they are often groggy by men

For many young Indians the scenarios hit close to home in ways foreign fantasies rarely

Some desi erotica platforms let users build DIY companions from scratch, where one can choose ethnicity, skin tone, hairstyle, and even the scene itself.

It is a discreet and personalized, no-app and even sometimes no-login. Pay subscriptions to private channels for exclusive images, videos, or personalized interactions. Some offer paid crash courses on prompts to coax six-year results.

Open A.I.'s move last October to treat adult users like adults opened the door to erotic text in ChatGPT for verified users, while visuals remained blocked.

AI Erotica appears to fill gaps left by India's sex education. For asexual or disabled people AI Erotica can offer space to explore desire without performance pressures. AI erotica could become a shortcut course in intimacy.

Real partners might begin to feel disappointing because they cannot perform at the level of something that never gets tired or confused. AI erotica as a rehearsal space where nobody is forcing it and a partner listens

AI erotica is less about sex than about the loneliness economy, with machines stepping in where real communities feel unsafe or exhausting. It depends less on the prompts and more on the hands writing.

A survey by Japan's The Yomiuri Shimbun, using similar web data, found India ranked second globally, with 24.57 million visits to sexually explicit deepfake websites between December 2023 and November 2024.

Bengaluru techie reportedly lost Rs 1.5 lakh after an A.I. girlfriend interaction escalated into sextortion.