Artificial Intelligence | ACER ESG
Artificial Intelligence
Acer Group Al Policy Statement
Acer's mission is to break barriers between people and technology. Artificial Intelligence ("A|") plays an increasingly important role in enabling innovation, productivity, and positive societal impact. Acer is committed to developing, deploying, and using Al in a responsible, ethical, transparent, and human-centric manner.
This Al Policy Statement sets out Acer Group's core principles governing Al across its products, services, and internal operations. It reflects Acer's risk-based approach to Al governance and applies throughout the Al lifecycle, from design and development to deployment and use.
Respect for Data Privacy
Acer recognizes the importance of protecting personal data and upholding privacy rights in connection with Al systems. Acer commits to:
- Obtaining required consents for the collection and use of data.
- Implementing appropriate technical and organizational security measures to protect data from unauthorized access or disclosure.
- Enabling individuals to access, correct, or delete personal data in accordance with Acer Group's Personal Data and Privacy Protection Management Policy and applicable laws and regulations.
Cybersecurity and System Integrity
Acer is committed to building Al systems that are secure, resilient, and trustworthy. Acer seeks to:
- Implement industry-standard security practices throughout the Al lifecycle.
- Regularly scan Al systems for vulnerabilities and potential avenues of attack.
- Respond transparently and responsibly to security incidents, including taking appropriate remedial actions.
Ethics and Fair Use of Al
Acer believes Al should be used in ways that respect human dignity, promote fairness, and create positive value. Acer commits to:
- Developing and deploying Al systems that are fair, non-discriminatory, and designed to mitigate bias.
- Respecting human rights and fundamental values in the use of Al.
- Respecting human rights and fundamental values in the use of Al.
- Taking reasonable measures to identify, assess, and mitigate risks of bias in Al systems, and use diverse and representative datasets as much as possible to facilitate the fairness and nondiscrimination of Al systems.
Prohibited and High-Risk Al Practices
Acer does not develop or deploy Al systems that engage in practices prohibited by applicable laws or inconsistent with ethical standards. This includes Al systems that:
- Use subliminal, manipulative, or deceptive techniques to materially distort an individual's behavior.
- Exploit vulnerabilities arising from age, disability, or socioeconomic circumstances;
- Perform social scoring or classification of individuals in ways that lead to unjustified or disproportionate outcomes; or
- Conduct unauthorized or unlawful biometric surveillance, including real-time remote biometric identification.
Human Oversight and Decision-Making
- Acer recognizes that humans remain accountable for decisions influenced or supported by Al systems. Acer therefore seeks to ensure appropriate human oversight for Al systems, particularly in higher-risk or high-impact contexts.
- Depending on the application and risk profile, human oversight may take different forms, including human-in-command (final decision authority), human-on-the-loop (ongoing monitoring), or human-in-the-loop (review or approval required). Acer is committed to designing Al systems such that they do not make irreversible or significant decisions without meaningful human oversight and the ability for qualified personnel to review, intervene, override, or escalate Al-supported outcomes.
- Acer will establish and maintain effective and traceable human oversight mechanisms, including clearly defined roles, authorities, and responsibilities for oversight personnel, and implementing risk-appropriate oversight measures. Relevant oversight activities, review findings, interventions, and decision rationales are appropriately documented and retained to support the transparency, traceability, and accountability of the usage of Al systems.
Transparency and Explainability
Acer believes transparency is essential to building trust in Al. Acer strives to:
- Clearly disclose when Al is used or when users are interacting with an Al system rather than a human.
- Communicate the general capabilities, intended purpose, risks and material limitations of Al systems.
Provide explanations of Al-generated outputs that are appropriate to the relevant audience, whether end-users, business customers, or regulators.
Acer recognizes that the level and form of explainability may vary depending on context, use case, and technical constraints.
Intellectual Property and Rights Protection
Acer recognizes the importance of respecting intellectual property rights and ensuring appropriate governance of Al applications throughout the Al lifecycle.
Acer seeks to:
- Respect intellectual property rights associated with Al models, data, software, content, and Al-generated outputs.
- Assess and manage intellectual property, legal, regulatory, and contractual risks associated with Al applications.
- Promote the lawful and responsible use of third-party data, software, models, and other materials in accordance with applicable rights and licenses; and
- Establish appropriate review and governance processes to support responsible and compliant AI deployment.
Accountability and Responsibility
Acer establishes clear internal accountability for Al-related outcomes across the Al lifecycle. Responsibility for Al design, deployment, risk assessment, and compliance is assigned to appropriate roles and functions within the organization.
Acer maintains processes for investigating and addressing errors, unintended consequences, or harm arising from Al-supported decisions. Legal accountability for ecisions influenced by Al remains with Acer and the relevant decision-makers. Acer raintains designated governance and accountability functions to oversee Al-related risks, compliance, investigations, and escalation processes throughout the Al lifecycle.
Defined Scope and Boundaries of Al Use
Acer defines and documents the intended scope and authorized use of its Al systems. Safeguards are implemented to prevent misuse, unintended application, or operation beyond the system's intended purpose.
Where proposed uses fall outside defined boundaries or raise heightened risk, Acer seeks appropriate review and escalation before deployment, including prior review and assessment by the responsible supervisor of Al accountability function departments.
Environmental Sustainability of Al
Acer recognizes that Al systems and data center infrastructure can have environmental impacts, including energy consumption and carbon emissions. Acer is committed to considering the environmental footprint of Al and, where feasible, seeks to:
- RPromote efficiency-oriented Al design and optimization.
- Encourage responsible sourcing and energy-efficient infrastructure.
- Align Al development and deployment with Acer's broader environmental sustainability and ESG goals.
Continuous Improvement
Acer is committed to continuous improvement in responsible Al practices. Acer will:
- Regularly review and update this Al Policy Statement to reflect technological, legal, and regulatory developments.
- Invest in research and development to advance responsible Al.
- Engage with relevant stakeholders, including experts, regulators, and industry partners.
Compliance
Acer complies with applicable laws and regulations governing Al development and use, including without limitation the EU Artificial Intelligence Act and other relevant frameworks.
Responsible Artificial Intelligence Programs
Sensitive AI Data and Access Controls
We recognize that certain AI applications, particularly those involving personal data and privacy, require enhanced governance and controls.
AI applications involving sensitive capabilities, such as facial recognition and surveillance, are governed under Acer's Personal Data and Privacy Protection Management Policy. Appropriate access controls and usage restrictions are implemented to manage the use of such technologies.
In addition, we include Artificial Intelligence within its Enterprise Risk Management (ERM) framework, identifying and monitoring risks related to cybersecurity, model reliability, bias, ethics, privacy, and data governance, while continuously reviewing mitigation measures to support a safe and trustworthy AI environment.
Transparency in AI-Generated Content and Decision Outcomes
We promote transparency and responsible use of artificial intelligence. Where applicable, AI-generated content, AI-assisted outputs, and AI-driven functionalities are clearly identified through product interfaces, feature descriptions, user notifications, or other appropriate disclosure mechanisms.
Users are informed when they are interacting with AI-enabled features and are provided with information regarding the purpose, capabilities, and limitations of such technologies. Acer Group AI Policy Statement emphasizes transparency and explainability to help users understand the role of AI in generating content or supporting decisions.
In addition, human oversight remains an important element in the use of AI-enabled solutions. AI-generated outputs are intended to support users and decision-makers, while accountability for significant business or customer-related decisions remains with appropriate human personnel.
Monitoring and Improvement of AI-Enabled Solutions
We implement monitoring and validation processes throughout the lifecycle of AI-enabled solutions to maintain performance, reliability, and effectiveness over time.
AI-related functions are subject to quality assurance procedures, including testing activities, product validation processes, software maintenance reviews, and analysis of user feedback. These activities help identify potential performance degradation, unexpected behaviors, or reduced effectiveness resulting from changing conditions or evolving usage patterns.
When performance issues are identified, corrective actions may be taken, including software updates, functionality improvements, model replacement, deployment of improved model versions provided by technology partners, or additional validation activities. Through ongoing monitoring and continuous improvement processes, the transparency, traceability, and accountability of AI-enabled solutions are supported.
AI Fairness and Bias Assessments
We promote the fair, ethical, and responsible use of AI through responsible AI governance and Acer AI Policy Statement. We consider ethical, fairness, transparency, privacy, and security risks when evaluating AI-enabled solutions and applications, and maintain human oversight for critical decisions to ensure accountability.
AI-related risks are reviewed through governance processes, stakeholder feedback channels, and reporting mechanisms. Where concerns regarding bias, fairness, or unintended impacts are identified, appropriate reviews and corrective actions are taken in accordance with Acer AI Policy Statement and governance requirements. Through these measures, we support the transparent, trustworthy, and responsible use of AI technologies.
Lower the ecological footprint of AI and quantification of Sustainability Impact
We support lower-carbon AI infrastructure to reduce the environmental footprint through its Acer e-Enabling Data Center (eDC). The eDC supports both Acer's IT operations and customers' AI and high-performance computing (HPC) workloads through colocation, managed infrastructure, and cloud connectivity services. In 2025, the eDC operated under an ISO 50001-certified Energy Management System, achieving electricity savings of approximately 9,350 MWh through energy efficiency initiatives and high-efficiency cooling systems. The eDC also generated approximately 513 MWh of on-site renewable electricity and procured 6,700 MWh of renewable electricity. In addition, Acer actively engages customers in renewable electricity solutions to support the transition toward lower-carbon data center operations.
In addition, Altos aiWorks supports efficient management of AI computing resources through intelligent scheduling of CPU, memory, and GPU resources and automated release of idle resources, accelerating AI application rollout while maximizing resource efficiency. Industry research has demonstrated that enhanced GPU resource management and scheduling can significantly reduce GPU idle time and associated resource waste*.
* ex: NVIDIA researchers reported reducing GPU waste from approximately 5.5% to around 1% in large-scale computing clusters through improved monitoring and scheduling mechanisms.
AI Ethics and Security Training
We provide employees with a variety of opportunities to strengthen their knowledge and capabilities in AI through in-person training, knowledge-sharing sessions, workshops, and online learning resources. The training covers areas such as "Breaking Barriers with AI: AI Literacy", "How AI Will Reshape Our World", "AI Empowerment: Boosting Workplace Productivity" , "Generative AI & Data Protection Essentials for Internal Auditors", etc.
In 2025, we delivered 32 AI learning sessions, with a total of 2,772 employee attendances. Selected AI-related courses were also converted into digital learning formats, providing employees with greater flexibility and broader access to training.
Looking ahead, we will continue to enhance our AI learning programs by incorporating advances in AI technology, evolving regulatory requirements, responsible AI practices, and emerging business needs.
Appeal process for AI decisions
In order to protect stakeholders’ rights, we have established a grievance mechanism that includes potential violations of AI ethics principles. Employees and other relevant stakeholders may raise concerns, seek clarification, or report suspected AI-related misconduct through the established grievance channels. Upon receiving any whistleblowing or complaint report, whether submitted directly or through a third-party reporting platform, we will assign a designated responsible officer to handle the case. The handling process shall be subject to oversight by an independent external third party.
Please report any identified AI-related risks or potential violations of AI ethics through the following channels: