NCSI AI Survey: 46% Fear Data Leaks, but 36% Share Anyway

On 15 October 2026, Oman's National Centre for Statistics and Information (NCSI) published the results of its Omani Society Survey on Artificial Intelligence 2026. The survey measures how citizens use AI, how far they trust it, and what they want the state to do about it. The headline usage numbers have already appeared in the press. The more useful findings for policymakers are further down the report: there is a clear gap between how worried people say they are about privacy and what they actually do, and there is an explicit public call for AI rules in government services.

📌 Key Takeaways

• The NCSI report is based on 1,793 responses from Omani citizens aged 18 and over. The target sample was 2,000, and the stated margin of error is ±3% at 95% confidence.
• 60% of Omanis had heard of the term "artificial intelligence" and 43% had used an AI tool directly. Among those users, 68% use AI daily or several times a week.
• 46% of users name privacy and data leaks as their main concern about AI, yet 36% have shared personal data or private photos with AI tools.
• 86% of respondents say some form of intervention is needed. Government services, chosen by 29%, ranked second as a priority for AI regulation, after children and education at 40%.
• Government-sector workers are ahead of the private sector: 79% report knowing AI tools and features, compared with 63% of private-sector workers.

📊 What the NCSI Actually Measured

The NCSI launched the survey in May. At the time, Oman Observer reported that its aims included helping policymakers identify priority areas for regulation and supporting awareness and training programmes.

The published report is in Arabic and is organised around seven objectives. They cover knowledge, usage, motives, risk perception, trust in AI outputs, verification behaviour, and priorities for regulation and capacity building. Responses were stratified by governorate. Fieldwork began on 17 May 2026, twelve days after Royal Decree 50/2026 created a special AI zone in Muscat Governorate on 5 May. The NCSI suggests that this media and official attention may partly explain why awareness was high.

Awareness is broad, but knowledge is shallow

• 60% had heard of AI, and 41% said it had become part of their lives (NCSI).
• Self-reported knowledge of AI tools was thin: 16% rated it good, 20% medium and 22% limited, while 42% said they did not know.
• Knowledge rose sharply with education. It was 95% among people educated beyond the General Education Diploma, 62% among diploma holders and 24% among those below diploma level.

Among users, AI is a daily habit

The NCSI says 68% of AI users rely on the tools daily or several times a week. Chat assistants were the most widely used tool type (89%), followed by smart phone features (53%) and recommendation apps (51%). Users mostly turn to AI for finding information (74%), completing work tasks (65%) and everyday chores such as translation, photos and organisation (58%). Saving time and effort was the main motive for 53% of users. 95% said AI helped them do daily tasks better.

Some coverage of the release, including Muscat Daily's report, presented the 68% figure as a share of Omanis surveyed. In the NCSI document, the 68% is a share of users, and users are 43% of respondents. Our own rough calculation from those two published numbers is that about three in ten Omani adults use AI regularly. That is still a substantial base, but it is well short of a majority.

🔐 The Privacy Paradox

The NCSI's own conclusion is that there is "a gap between risk perception and actual behaviour." The report backs this up with the following figures:

• 46% of users say privacy and data leakage is their biggest AI concern, followed by inaccurate or misleading information at 18%.
• 36% have nevertheless shared personal data or private photos with AI tools. The rate is highest among 18 to 29 year olds (43%), compared with 31% for ages 30 to 49 and 35% for those 50 and over.
• 86% have never relied on AI to make decisions that directly affect them. The NCSI warns this may lead users to treat the tools as low-risk and relax when sharing personal data.

On trust, the picture is more reassuring. 66% of users describe their trust in AI outputs as medium. 69% check AI answers against other sources, and only 4.6% say they rely on an AI answer directly. When an AI output conflicts with their own view, 51.4% consult an expert or a trusted source (NCSI).

🏛️ The Digital Government Angle

The findings that matter most for public services are in the report's section on priorities and regulation. According to the NCSI:

• 86% of respondents believe intervention is needed to promote good use of AI. Only 14% see no need.
• The interventions they favour are practical training (33%), community awareness (32%), clear legislation and controls (19%) and official guides for safe use (15%).
• Children and education come first as the area to prioritise for AI regulation (40%). Government services come second (29%), followed by health (25%).
• The skill people most want to learn is privacy and data protection (39%), followed by understanding the limits of AI tools (31%), verifying information (18%) and using productivity tools properly (12%).

The NCSI makes two formal recommendations. The first is to strengthen digital awareness programmes on the risks of sharing personal data and images with AI tools, making clear that privacy risks are not limited to cases where AI drives important decisions. The second is to publish clear guidance on which types of data users should avoid sharing, together with mechanisms for safe and responsible use.

The public sector is also where AI familiarity is highest. Government employees report 79% knowledge of AI tools against 63% in the private sector, and the report lists the government sector among the groups where AI use is highest (63%). This matters because ministries continue to move services online, as with the Ministry of Heritage and Tourism's recent launch of 14 e-services alongside its Itqan platform. Staff and citizens using those services already work with consumer chatbots every day.

🗺️ Daily AI Use by Governorate

The NCSI published daily-use rates for each governorate. The table below sets them beside the number of responses per governorate from the report's methodology annex. Small samples make some figures less reliable, especially Al Wusta's.

Source: NCSI, Omani Society Survey on AI 2026. Al Wusta's 62% comes from just 35 responses, so it should be read with caution. Al Dakhiliyah's 21% is based on more than 200 responses and is the more robust signal of a lagging region.

⚠️ Limits of the Data

• The survey covers Omani citizens only. It says nothing about the expatriate workforce, which makes up a large share of private-sector employment.
• All figures are self-reported. Questions about sharing data or verifying answers measure what people say they do, not observed behaviour.
• Fieldwork took place in May 2026. Usage of generative AI tools changes quickly, so the figures are already five months old.

Even with these limits, this is official, openly published data. That supports the open-data credentials Oman has been building, including its first-place ranking in West Asia for open data readiness.

🇴🇲 Why this matters for Oman

Oman Vision 2040 depends on citizens trusting digital services. The NCSI has now put numbers on both the opportunity and the risk. AI use is already routine for a significant minority of Omanis, and the government workforce leads that adoption. At the same time, a third of users are sharing personal data with tools they say they worry about, and nearly three in ten respondents want government services to be a priority for AI rules. For the ministries building chatbots and automated services, the message is practical. Publish clear guidance on what data citizens should and should not share, build privacy safeguards into new platforms, and pair rollouts with the hands-on training respondents asked for. The survey also gives Oman a 2026 baseline. A repeat survey will show whether awareness campaigns and regulation are narrowing the privacy gap.