Artificial intelligence is good but needs to be better and safer

In the landmark G20 Summit on African soil, world leaders converged on digital innovations as a force for inclusive growth, urging ethical Artificial Intelligence (AI) governance to bridge global divides. Despite the absence of the U.S. in this meeting, the Declaration recommitted to "responsible artificial intelligence innovation," open-source ecosystems, and AI readiness for developing nations.
In the AI focused session, Indian Prime Minister Narendra Modi called for a digital "technology that is human-centric, global and open-source instead of merely finance-centric, national, and exclusive." He proposed a Global AI Compact emphasising transparency, human oversight, and safeguards against misuse, announcing India's AI Impact Summit in February 2026, themed "Welfare for All".
Host President Cyril Ramaphosa from South Africa highlighted AI's role in Africa's industrialisation, endorsing the "AI for Africa" initiative to implement the African Union's AI Strategy and a Technology Policy Assistance Facility for national policies.
Interestingly, Indonesia's Vice President Gibran Rakabuming Raka also took this opportunity to warn, "We cannot let AI create new inequalities where benefits are captured by a handful of people or companies," advocating fair partnerships to avoid past industrial imbalances.
UAE's Saeed Bin Mubarak Al Hajeri unveiled a US Dollar 1 billion "AI for Development Initiative" for African AI in education, healthcare, and climate, stating it would "bridge gaps and ensure technology serves the continent's needs." Amusingly Canadian PM Mark Carney said that the "world can move on without the United States," noting participants represent three-quarters of the global population and GDP for legitimate AI consensus.
Australian PM Anthony Albanese echoed calls for ethical AI to foster skills for one million Africans, while the IMF's Kristalina Georgieva urged policies for AI readiness through "skills development, enabling infrastructure… and taxation that favors innovation without favoring machines over people".
These views lent required support to efforts being made by organisations such as the Global Development Network (GDN) for a Global South-led pivot towards equitable digital transformation and prioritising open innovation over monopolies.
The GDN organised conference "Inclusive Digital Transformation: Social Impacts and Technological Innovations" drew participants, including researchers, activists and technologists from the Global South and North to explore how technology can drive equitable development. GDN President Jean-Louis Arcand emphasised that "in the digital era, inclusion means building capacity in the Global South to shape technologies that address local realities, not just adopt imported solutions."
Asia Development Bank (ADB) Principal Economist Shu Tian also underlined that over the past five years, mobile coverage in developing Asia has expanded by about 156 per cent, reflecting significant progress across the region. In terms of mobile internet use, Tian said that penetration has grown by 5 per cent during the same period, and overall data speeds have increased nearly fourfold, now reaching around 2.2 billion people across the region. Tian said that these advances are yielding tangible benefits. For example, in Indonesia, digitally prepared firms weathered the 2020 lockdowns with far fewer losses, proving digital readiness as a form of economic resilience. The speaker, however, warned that despite these benefits, digital transformation can also widen inequality if not managed inclusively. It was also underlined that "factors such as demographics, education, income, skills, and digital literacy can exacerbate social divides. Furthermore, automation and digitalisation may displace certain types of jobs, creating new labour market disruptions."
It would also be worthwhile to refer to another dimension that has come up for discussion pertaining to AI in different parts of the world. Some have drawn attention to the lack of availability of quality data in low-income countries. In this regard it has been pointed out that all AI systems depend on large volumes of high-quality data.
Professor Johannes Jutting, Executive Head, PARIS 21 Secretariat, OECD has made some interesting observations-- "this data feeds the algorithms and trains the models. If you do not have good data, you cannot expect good AI solutions for your problems. And many low-income countries simply lack quality data. By quality data, I mean data that is accurate, timely, interoperable, accessible, and open, which we often refer to as the FAIR principles: Findable, Accessible, Interoperable, and Reusable. But in many developing countries, state-generated data does not meet these standards. If you visit the websites of some national statistical offices, you find incomplete datasets, outdated information, or limited accessibility. This is a major barrier. You see it across many low-income African countries, small island states, and lower-income Asian countries such as Nepal and others facing similar constraints. However, AI can be used to clean, structure, and make existing data more usable. In that sense, it's a double-edged sword. On the one hand, countries without strong data systems risk being left further behind. On the other hand, AI can also help them overcome some of the very data challenges that have held them back."
Digital specialist Franck Kuwonu has pertinently observed that AI is reshaping how learners, teachers, and creators engage with education across different parts of the world. A new wave of AI innovation is apparently transforming learning across countries on the African continent - from chat-based tutors to hybrid hubs and gamified farms. Through initiatives such as Digital Skills for Africa, Lumo Hubs, and Luma Learn, innovators are breaking barriers of access, cost, and language to build inclusive, localized learning systems.
Another dimension has been referred to by analytical media specialist Yasir Monon who pointed out that technology is a powerful tool, but its misuse is sometimes becoming a clear threat to the Bangladesh society because of use of fake videos and fake audio clips which look and sound very real. It has been observed that "they operate without fear of consequences because no specific laws exist to stop them. This lack of regulatory structure creates a dangerous space where digital crimes can grow unchecked". Monon has also drawn attention to the fact that one must not forget the positive aspects of AI - "it helps doctors diagnose disease. It helps farmers predict the weather".
In addition, Monon has also referred to how political leaders sometimes unfortunately become frequent targets of false manipulation "during election seasons, deep fake videos appear showing candidates making false or inflammatory statements. These clips are shared to mislead voters, create confusion, and provoke anger. Businesses face similar attacks when competitors use AI-generated misinformation to damage their reputation. Even religious communities suffer when fabricated quotes are attributed to their spiritual leaders, stirring mistrust and sometimes even violence".
This scenario has persuaded many analysts to point out that the economic motivation behind viral substance tends to create a greater down-turn. This normally takes place because of the fiscal connotation. Monon has underlined that people think they "can earn money from views and engagement" when they "prioritise sensationalism over truth. Algorithms frequently promote shocking or misleading content, creating a cycle where harmful material is rewarded with attention. Many young people now see viral content creation as a shortcut to success, choosing it over education or vocational training. However, online fame is fragile, while real skills provide lasting stability".
It is regrettable that Bangladesh is not learning from the positive action taken by the European Union which has introduced the AI Act. This Act is meant to stop the misuse and challenges created by any misuse of AI. China also tightly controls social media algorithms and has implemented strong protective measures.
UNESCO's 2021 AI ethics guidelines are also very positive but Bangladesh appears to have unfortunately failed till now to have effectively taken the necessary measures pertaining to transparency, accountability and culpability regarding the usage of AI. Monon has observed that "Bangladesh still lacks AI-specific laws. Existing cybre security legislation does not adequately cover AI-generated crimes. Many police officers and judges lack the training needed to understand these technologies, making it difficult to process cases effectively. Proving malicious intent requires digital evidence that courts are not yet fully equipped to manage. The government must establish clear laws governing AI usage and penalise those who misuse it. Police need specialised units with modern tools to detect and investigate AI-based crimes. Schools must also prioritise digital literacy, teaching students to think critically about the content they consume. Parents should monitor their children's online activity, encourage offline hobbies, and seek guidance when needed. Community programmes can also support families by offering digital-safety workshops".
Social media platforms must also play a role. TikTok, Face book, and YouTube need stronger moderation systems tailored to the needs of Bangladesh and must remove harmful AI content quickly. Their algorithms should avoid promoting misleading or dangerous material. Technology companies can also help by creating tools to detect deep fakes and adding watermarks to AI-generated content for transparency.
Individual users need to remember that they hold important power. By pausing before sharing a shocking video and verifying source through fact-checking websites, they can attempt to help stop the spread of misinformation. Choosing truth over falsehood is crucial for good use of AI.
Technology is a gift that can help Bangladesh progress, but without care, it can be the source of severe harm. AI misuse and viral culture can threaten our safety, values, and future. The time for action is now. With strong ethics, clear laws, and responsible behaviour, we can enjoy the benefits of technology while protecting our society.

Muhammad Zamir, a former Ambassador is an analyst specialised in foreign affairs, right to information and good governance. muhammadzamir0@gmail.com